encounters a java.lang.NullPointerException

when resuming replication after changing the

maintenance user id.

Customer have installed "rs_install_statreport_v1510_ase.sql".

When Customer populate statistics using "rs_stat_popultae", this store proc call "rs_stat_populate_netspd" which report duplicate key error:

1> rs_stat_populate

2> go

We are going to populate report. Script version 1510. :-)

Msg 2601, Level 14, State 2:

Server 'SYBASE_MEDISREP2', Procedure 'rs_stat_populate_netspd', Line 189:

Attempt to insert duplicate key row in object 'rs_statreport' with unique index 'rs_key_statreport'

Command has been aborted.

...

Because Customer have setting up a Warm/Standby configuration after he have create the Primary (Active) Connection, the Logical-Connection have the same name ("dsname" and same "dbname") but as well the same Database-ID ("dbid).

In the following example, see Connection-ID 144, but as well as Connection-ID 170:

1> select dbid, dsname, dbname, ltype, ptype, ldbid from rs_databases order by dbid

2> go

dbid dsname dbname ltype ptype ldbid

-------- --------------------------- --------------------------- ----- ----- --------

101 SYBASE_MEDISREP2 rssd P A 101

144 SYBASE_MEDISPRD1 facets221ar P A 144

144 SYBASE_MEDISPRD1 facets221ar L L 144

145 SYBASE_MEDISREP2 facets221ar P S 144

170 SYBASE_MEDISPRD1 facets221 P A 170

170 SYBASE_MEDISPRD1 facets221 L L 170

171 SYBASE_MEDISREP2 facets221 P S 170

(7 rows affected)

Such configuration cause "rs_stat_populate" to report "Attempt to insert duplicate key row in object 'rs_statreport' with unique index 'rs_key_statreport'".

DSI (connection) levels. This property controls when newline characters, contained as a part of the

data, are escaped in order to preserve them. Originally, these characters were escaped such that

replicate databases similar to Sybase ASE would not remove them. This behavior resulted in

duplicated newline characters in some non-ASE databases.

"dsi_newline_escape" valid values are "on"|"off".

The default value is "on" to match current Replication Server behavior.

"dsi_newline_escape" may be set through the following Replication Server syntax:

configure replication server set dsi_newline_escape to ["on"|"off"]

alter connection to <connection> set dsi_newline_escape to ["on"|"off"]

For replication to MS SQL and ASE, dsi_newline_escape should be "on". For replication to Oracle and

DB2 LAN, dsi_newline_escape should be "off". This new property has not been verified for any other

replication targets.

1> rs_stat_populate

2> go

We are going to populate report. Script version 1510. :-)

Msg 247, Level 16, State 1:

Server 'PRIM', Procedure 'rs_stat_populate_sqt', Line 260:

Arithmetic overflow during implicit conversion of NUMERIC value '4.524371180900952110274264601392638908625' to a NUMERIC field .

(return status = -6)

=> Tested using different RS releases including RS 15.6 "rs_install_statreport_v1510_ase.sql" script.

=> This issue is different that CR #609448, WA of CR #609448 was already implemented during this testing and the error message is different.

=> ASE hosting RSSD does not hit ASE CR #579925:

1> select @@version

2> go

-----------------------------------------------------------------------------------------------------------------------------------

Adaptive Server Enterprise/15.5/EBF 18160 SMP ESD#2/P/Sun_svr4/OS 5.8/asear155/2514/64-bit/FBO/Wed Aug 25 16:52:39 2010

1> select (-8646682.3 - 2078078.4 - 177488.80) as AAA

2> go

AAA

-------------

-10902249.50

Customer is hitting RepServer error #8039:

E. 2011/09/02 17:41:29. ERROR #8039 SQM(1126:0 LON_GDS_PROD1.gds_rep) - tr/mdext.c(2009)

MD failed to wake someone waiting for memory from source=LON_GDS_PROD5.gds_live.

As Customer claim this does not make SQM thread going down, Customer would like this error been reported as only a "Warning" instead of as an "Error" (Customer is having some scanning tool that look for "ERROR" keyword and ping DBA)

1) connect to a repository

2) check-in project.pdm model

(check in options - 'check out after check in' and 'save diagram for Portal' are selected)

3) model checks in correctly and diagram is visible in portal

4) in PD, add new table to the diagram

5) check in model

5a) uncheck 'check out after check in' - by default 'save diagram for Portal' is disabled

5b) click simplified

5c) click advanced

5d) click 'save diagram for Portal' but do not select 'check out after check in'

5e) click OK to check in the model

Messages state that the diagram has been saved.

However, in the Portal, the diagram is updated with the new objects, but the html links for the new objects are not created.

- Power Designer 15.3 EBF#10

- Oracle 10g Repository

- Native Connection Profile for Oracle 10g

Once a model in repository is locked (even if performed by ADMIN user), no user is able to unlock it any longer, if using Oracle Native Connection Profile.

- Power Designer 15.3 EBF#10

- Oracle 10g Repository

- Connection Profile Native connection Oracle 10g

Permission settings not being saved in 15.3 EBF#10 with Oracle native profile. If using ODBC Connection instead, it works. This is a regression from previous PD 15.3 EBF#9, where native Oracle connection profile worked ok.

Customer have configured his RepServer using Embedded RSSD (eRSSD) and is starting RepServer instance as a Windows service.

RepServer service is started using "Local Account".

When requesting an eRSSD backup using "admin erssd, backup" command or when RepServer issue automatically the backup, the backup process hang and lead eRSSD un-accessible (ASE engine is hanging) and eRSSD transaction log is not truncated.

This cause RepServer not able to access to its RSSD:

E. 2011/09/26 17:42:45. ERROR #11060 USER(sa) - s\stscol.c(2079)

CT/CS Lib function 'ct_results' failed. Retcode = 0.

E. 2011/09/26 17:42:45. ERROR #1027 USER(sa) - s\stscol.c(2079)

Open Client Client-Library error: Error: 84083974, Severity 5 -- 'ct_results(): network packet layer: internal net library error: Net-Library operation terminated due to disconnect'.

I. 2011/09/26 17:42:48. Command failed: D:\sybase\REP-15_2\ASA11\win32\dbping.exe -q -c "eng=SYB_PRD1_RS_ASA;Language=us_english;CommLinks=tcpip(DOBROAD=NONE;HOST=localhost;PORT=8810)". Operating System returned: Unknown system error '0'

I. 2011/09/26 17:42:48. Starting embedded RSSD: D:\sybase\REP-15_2\ASA11\win32\dbspawn.exe -f -q D:\sybase\REP-15_2\ASA11\win32\dbsrv11.exe -ti 0 -x "tcpip(PORT=8810;DOBROAD=NO;BLISTENER=NO)" -o D:\sybase\REP-15_2\install\SYB_PRD1_RS_ASA.out D:\SYB_PRD1_RS_ASA\SYB_PRD1_RS_ASA.db

This may be due to some permission issue on eRSSD backup-ed file, even Windows setting report correct permission setting.

But, RepServer or ASA "dbbackup" may report an error message or warning message instead of way indefinitely or hanging.

It only supports 9 and 10.

But users can select Oracle 8 and 11.

When updating an existing repository, we need to change all PK indexes to unique clustered index.

duplex online redo log, Replication Agent may

go into the REPLICATION DOWN state due to a

log scanner error stating 'Invalid record

offset with block id <XXX,XXX> and block

offset <XXX> specific on log device <XXX>'

after switching from one redo log member to

another.

OdbcError: [S1000] [Oracle][ODBC][Ora]ORA-12899: value too large for column "SYSTEM"."PMXATV"."VALE" (actual: 256, maximum: 254)

conneting as "normal" user (without "Manage all Documents" rights)

-> can not see any documents in Repository Browser

Worked fine with PD 15.3 Viewer

customer is asking for a "find object" feature to be added, so that the users who view diagrams through PD Portal can easily locate the objects in the diagram, especially if the diagram is large

a log operation sorting error when the Oracle

System Change Number(SCN) rolls over the

n * 0xffffffff boundary.

The page that customer is referring to can be accessed via

1) Log into Portal

2) Click Users and Groups from Portal menu

3) Click Groups from the the tree on left pane

4) Click on a Group name (i.e. ADMN)

5) Click on Members tab within the Group

6) The Members pane only shows Name (login) and Realm

for LOB that is not enabled for replication.

PD 15.2 EBF#11

MS SQL Server 2008 as backend

PD using Native SQL connection to Repository (SQL Native Client)

This is a consult CR from Itau-Unibanco

Currently, customer delays around 50 seconds to connect to Repository. We confirmed all patches, indexes recreations etc. (from previous cases) were applied successfully into Repository.

Customer is asking if there's any extra recommendations to improve this connection time to repository.

PD 15.2 EBF#11

MS SQL Server 2008 as backend

PD using Native SQL connection to Repository (SQL Native Client)

This is a consult CR from Itau-Unibanco

Currently, customer delays around 50 seconds to connect to Repository. We confirmed all patches, indexes recreations etc. (from previous cases) were applied successfully into Repository.

Customer is asking if there are any extra recommendations to improve this connection time to repository.

Replication Server - Real Time Loading Edition 15.6 and Replication Server Heterogeneous Edition 15.6.

see the select statement

A0157 04utf8distribute 2 ~"!,3 5 sqlddl append first last ~#"Zselect s

et_enc_password(0xc007caaa89d95f60fd20ac92a3c03206e48e9aeb7912465b7aafb

752babaa42de30bc66af414bc2817c8, 'jstask')

Replication Agent 15.7

for Oracle allows user

to mark table contains UROWID

column through 'force' option,

but UROWID column itself

will always be filtered.

This functionality is also

merged to 15.6 and 15.2.

configuration parameter lr_ntext_byte_order.

New values of Position, Relation Position and Size have not been saved.

Compare local model with Repository keeps detecting differences.

Prevention for

1)Oracle compressed table

prior to Oracle 11gR2;

2)Oracle encrypted column and

tablesapce prior to Oracle

11gR1;

3)Oracle secure file LOB prior

to Oracle 11gR2;

4)DB2 row compressed table prior

to DB2 v9.7.

Repository can not be shown, can not create user.

Setup repository with other databases, PMTEMP, PMTEMP2... are not temporary table, it works OK.

Update a document version, 'Check out after check in' does not check out recent objects so the local version is not up-to-date.

This does not happen if custom Comparison Options are not set by default.

2a) Issue "repserver -upgr" will upgrade Replication Server at the same time or

2b) Issue "sysadmin upgrade,'database'" after upgrading Replication Server with rs_init or repserver with the -nodb option.

giving the following situation (with replicate minimal column and groupid is not a primary key column on the child table):

parent table:

insert into groupid values (1)

insert into groupid values (2)

Child table (groupid column on child table has RI on the groupid table)

update locid set groupid = 2 where groupid = 1

Parent table:

delete groupid = 1

This will work at the primnary. However, with hvar, update locid will continue to be a update due to minimal columns and the fact that groupid is not named as a primary key column. Since HVAR always applies deletes before updates, the delete fails with constraint violation, which therefore makes the entire transaction to be applied by language.

ORACLE ONLY: Replication Agent sent

duplicated locator to Replication server

when handle insert with LOB/long data type.

Provide new ra_downgrade and enhanced

ra_migrage API to improve upgrade

and downgrade process.

E. 2012/02/16 13:50:06. ERROR #2100 REP AGENT(p6x155e5sdx.pdb1) - s/packer.c(8434)Packer failed with bad multibyte string data. Length=57(bytes). String='Xxxxxx Xxxxxxx Xxxxx- xxxxxx xxxxxxx xxx...'.

when ASE -> RS -> IQ ... with ...

1) 64bit ASE 155e5 pds.pdb on linux

use sqllocres to change the default code page to utf8 and sort_order to bin_utf8

sqlsrv.default_characterset: utf8

sqlsrv.sort_order: bin_utf8

2) IQ 15.4 on linux using ... collation 'UTF8'

create database

...

case respect collation 'UTF8' java off

...

go

3) multibyte data bcp'd into ASE pds.pdb

remains in REPLICATING state but fails to

replicate any changes after processing an 'alter

index ... rebuild' log operation.

There are several samples where a wrong or obsolete path structure of models using named path are not possibe to be corrected via the PowerDesigner GUI.

The path in the PMFLDR (XNAM) or PMXFIL (XNAM)persist in the repositroy.

During a check in, the subpath information can not be modified.

It persist in the repository no matter how you try to resolve it.

- Power Designer 15.2 EBF#10

- Windows XP 32-bits

- SQL 2005 and SQL 2008

Problem Description: Repository is displaying duplicated folders in the repository, when there are more than one branch with the same model.

to what is reported by RS "admin stats, mem_in_use" command.

The issue is Solaris memory management does not release memory used by a process immediately once it is allocated even though Rep Server already freed the memory by "free()".

In effect, "prstat" represents maximum RS memory use.

A new RS command is provided in this resolution: "admin stats, max_mem_use" to get the max memory space used since RS start up. This number should be comparable to the Solaris "prstat".

A error message is always being thrown out:

---------------------------

PowerDesigner - Error

---------------------------

You may not connect to the repository because your account has been inactive for more than 365 days. Contact your administrator to unblock your account.

1. Setup the proxy for instance to a MS SQL or ASA database where repository 16.1 is always is installed

2. Setup proxy repository definition using the proxy connection

Repository -> Repository Definitions ...

3. Login as Administrator to the repository in Repository Browser

4. Create a new user "Stefan"

Repository -> Administration -> Users

5. Press the "Reset Password" button and set the password to "Sybase001"

6. Disconnect from Repository

7. Connect as Stefan to the Repository using the Proxy

-> PowerDesigner - Error

value of 16K is too small, frequently resulting

in DB2 SQL error 2650 and reason code 8. The

default value is now 64K.

clarify:

1. When manual says that "Replication Server only uses LDAP for ...server names and connection properties". The result is you may use and LDAP server in place of an

Interfaces file. LDAP is only supported for server/interfaces access.

2. You can NOT use LDAP with Replication Server for Replication Server user

authentication. Further explanation that normally *ONLY* RS admins need to log in to the RS -- this is not a system that has or should expect a lot of logins from general user population. More often, monitoring processes NOT LIKELY in LDAP will be the more normal user logins (via RMS/UAF/SCC). Accordingling, LDAP user authentication has limited affect/impact within any organization and NOT supported by Replication Server.

But dsi_xact_group_size can be used on RS 15.7.

ORACLE ONLY: Oracle11g XMLType table

replication feature support.

consistent ... the password arguments should function the same for all three of

these utilities ... rs_subcmp is different compared to isql and bcp. At least

isql and bcp are consistent.

When last line and last character of (=) in the cfg file ends with ... PPWD =<absence of a carriage return line feed characters by the way> ...

PPWD =

then the following message happens ...

Warning: line 7 in configuration file does not contain a valid assignment. [\subinit.c(600)]

this only happens on windows

migration, the Replication Agent will prompt the

user that migration has succeeded. But the

Replication Agent tables in the primary database

may not have actually been upgraded or downgraded.

- PD 16.1 GA and 16.1 EBF#2

- Repository DB: Microsoft SQL Server 2008

- Using PD 16.1 Proxy (However, issue also occurrs from outside PD proxy)

When checking in a model in PD 16.1, customer receives the following errors:

Message box: "Unexpected error while executing SQL statements in repository"

If using PD proxy, customer may also receive the following message box:

PDProxy16.exe application error:

"The exception unknown software exception (0xc0000417) occurred in the application at location 0x78b2a3f6"

And PDProxy16.exe error log shows data mismatch as root cause of check in:

2012-03-16 10:26:59 info - OnClose: Closing client connection 10.2.209.119

2012-03-16 10:26:59 debug - SSockClient: closing #3

2012-03-16 10:27:00 debug - status: 0 clients, 2 sockets, 1 db connections, 0 for pipes, 5 own/882 total threads, 168.1MB private bytes, 646.5MB available of 2.1GB physical memory, 2.7GB available of 4.1GB virtual memory

-------------------------------- 16.1.0.3637 Unicode

2012-03-16 10:27:06 debug - CreateSocket: Listening on created socket.

-------------------------------- 16.1.0.3637 Unicode

2012-03-16 10:33:57 fatal - SProxyQueryColumn: unsupported length for SRepCHAR 0

2012-03-16 10:33:57 fatal - SProxyQueryColumn: unsupported length for SRepCHAR 0

2012-03-16 10:34:23 fatal - SProxyQueryColumn: unsupported length for SRepCHAR 0

2012-03-16 10:34:31 fatal - SProxyQueryColumn: unsupported length for SRepCHAR 0

2012-03-16 10:34:31 fatal - SProxyQueryColumn: unsupported length for SRepCHAR 0

2012-03-16 10:38:07 error - OdbcError: [22026] [Microsoft][SQL Server Native Client 10.0]String data, length mismatch

And on PD Console you will see:

Checking in the document 'C:\Cases\11726366\DBQLADM.pdm'

- Saving the local document in the repository...

Unexpected error while executing SQL statements in the repository

The check in of the document has been cancelled.

alter database replication definition db_repdef

with primary at PDS.PDB

not replicate tables in ('owner1.a*')

ORACLE ONLY: In a Multiple Replication Agent

environment, when we run 'pdb_xlog init' again

after running 'pdb_xlog init' and 'pdb_xlog remove',

the following error comes along.

Could not read upgrade finialization status on

primary database <"orcl">, because <null>.

This situation occurs if remote store procedure raise a user-error and also return a result-set after having raised the user-error:

- if the user-error is assigned (in RS side) to STOP_REPLICATION, DSI goes down on user-error (=> this is expected behavior)

- if the user-error is assigned to IGNORE, DSI goes down with 5216

1) When the Repserver's site_version is at 1570, lists the replication definitions which exist to only define primary key columns and/or quoted table/column names.

2) When the Repserver's site_version is at or greater than 1571, lists the repdefs which exist to only define primary key columns, quoted table/column names and/or customized function strings.

remote database if the archive path is not set.

LogMiner configuration/permission when

replication starts rather than during

Replication Agent initialization.

'ra_locator zero' command, the trunction point is

positioned at the end of the log instead of at the

beginning.

Repository: MS-SQL Server 2008

This is for Itau-Unibanco.

Some indexes from repository tables were not dropped/recreated properly during repository migration process.

Customer needs to know if the same 15.2 scripts they used to drop/recreate the indexes can also be safely used in production repository. On QA environment, they used the scripts and repository connection time/performance was very good after that.

Checking in the document 'XXX'

- Saving the local document in the repository...

Unexpected error while executing SQL statements in the repository

A command sent to the repository proxy failed.

The check in of the document has been cancelled.

Working well via direct ODBC connect.

execution fails due to trailing whitespaces in

the parameter values.

archive logs too soon, before Oracle completes

its archive work.

parameter but does not restart the Replication

Agent, and then execute 'ra_admin init', the error

'Could not initialize the primary database

because:ORA-00942: table or view does not exist'

comes along.

fails if the user specified by ra_admin_owner is

different than the user specified by

pds_username.

Replication Agent instance and issues the 'resume'

command, an error might occur: "Oracle LogMiner

must be installed in order to use this command.

Please verify that LogMiner is installed."

property pdb_thread_filter can be set but has

no affect.

When repository administrator is updating a model into the repository, the model is saved with wrong version number and another versions with totally different version numbers are created at the same time. Only one of these versions contains check in comment.

If user does not manually delete the unexpected versions, he is not allowed to update the model anymore.

in compiled update/ or delete commands replication (through join-update/delete operations).

DDL that is executed by SYS user.

and replicated when they should not be.

sometimes sends an inaccurate length attribute

for an image column to Replication Server

resulting in data loss.

to reproduce:

1) connect to portal in IE 8 or 9

2) via repository explorer, click on an object to 'view detailed information' (diagram or text)

3) select Print Preview

NOTE: The header, footer, and explorer treeview are visible. However, the details of the selected object is not included in the printout

Invalid object identifier for table or function 'tab1_repdef'. id = 0x0102006500000065.

E. 2012/04/27 13:34:37. ERROR #5035 DSI EXEC(107(1) ost_repsolvm_01.rdb2) - /dsiutil.c(4091)

Error from unpacker or parser. See previous message for more information.

to replicate IMAGE column data.

failed to be replicated by Replication Agent.

procedure when the procedure owner (schema)

is different than the pds_username user.

LogMiner fails to start with a 0 start SCN.

no metadata found in marked object.

VirtualFileNotFoundExceptions sends Replication

Agent to REPLICATION DOWN state.

NoSuchElementException error when processing a rollback operation.

goes to REPLICATION DOWN state when replicating

SecureFile LOB.

to the RS server, then when a warm standby switch is initiated with the

execution of the RS command 'switch active', the warm standby switch

fails to complete.

Repository on MS SQL Server 2008

Issue description:

Our customer is now using PD 16.1 EBF#4. He's using LDAP authentication for the repository connection. They are not using SSL. The issue is: on PD Portal, he connects fine. On PD itself, the connection fails

Customer's text:

In the repository we must to be able to create user VIEW following different perspective.

Currently PowerDesigner allows to group objects by Branches and Folders and present them in that form.

It would be advantageous to keep the actual physical structure but also provide the ability to group objects and display objects in other VIEWS. We would like create different perspective (outlook, context), examples :

view "List of objects associated with a user"

view "List of conceptual models"

View "List of projects having working on this subject",

etc.

replicate a rollback.

ERROR #50 DSI EXEC(700(1) SYBOLTPWS.VBDMASTR) - /nrm/nrm.c(733)

'rs_objects' system table is corrupt."

� dsi_num_threads to 5

� dsi_num_large_xact_threads to 2

� dsi_serialization_method to 搘ait_for_commit�

� dsi_sqt_max_cache_size to 1 million bytes (on 32-bit platform)and 20 million bytes (on 64-bit platform).

With wait_for_commit, DSI will send the next transaction when the previous transaction is ready to commit, but has not finished commit, it may cause data inconsistency at replicate site. An example is:

Step 1. thread 1: begin tran, insert into table A

Step 2. thread 2: begin tran, delete the row thread 1 just inserted

Step 3. thread 1: commit

Step 4. thread 2: commit

The delete at step 2 will affect 0 rows, since the insert at step 1 is not committed yet, leaving the row at replicate database.

with unique index 'rs_key_profdetail'" and upgrade fails when you try to upgrade Replication Server 15.7 to 15.7.1.

PD is currently in a lab environment. The Proxy and SQL Server database are on different hosts. The Proxy is on a VM and the SQL Server instanced is on a shared host.

In preparation, they began to perform load and stress testing with the repository and have run into the issue that follows.

During a load and stress testing exercise, most models check-in and checkout without issue. However, when performing multiple check-in\outs with their large models, they receive the following errors at different points.

fatal - Exception caught in OnReceive

OdbcError: [S1000] [Microsoft][ODBC SQL Server Driver]Connection is busy with results for another hstmt

OdbcError: [S1000] [Microsoft][ODBC SQL Server Driver]Unknown token received from SQL Server

OdbcError: [S1000] [Microsoft][ODBC SQL Server Driver]Unspecified error occurred on SQL Server. Connection may have been terminated by the server.

OdbcError: [01000] [Microsoft][ODBC SQL Server Driver][SQL Server]The statement has been terminated.

OdbcError: [08S01] [Microsoft][ODBC SQL Server Driver][TCP/IP Sockets]General network error. Check your network documentation.

OdbcError: [24000] [Microsoft][ODBC SQL Server Driver]Invalid cursor state

OdbcError: [08S01] [Microsoft][ODBC SQL Server Driver]Communication link failure

NOTE: the proxy log includes information for multiple versions of the proxy.

When these errors occur, they encounter deadlocks on SQL Server.

initialized, issuing 'pdb_xlog deinit, force'

returns success when an error message is expected.

down with a NoSuchElementException error when

replicating LOB data.

- PDF export or

- print function which prints the diagram as whole

( which may result in PDF also )

increases unexpectedly because Replication Agent

generates three LTL commands for a single update

to an Oracle table with a column of type Object.

deadlocks on the Replication Agent DDL table at

the primary database when more than one DDL

operation is executed in parallel.

In an environment with only one DSI executor thread, the current group algorithm with SQT cache size as a parameter has a defect, which generates single transaction group.

upgrade/downgrade are missing from the Replication

Agent installation.

unresponsive after execution of the

pdb_get_columns command.

drop a non-existent replication definition for a

table or procedure is misleading.

connects to Replication Agent is 15.5 or

later, Replication Agent no longer addes

the 'replicate minimal columns' clause when

creating replication definitions.

after an in place upgrade from earlier versions

to 15.7.1 GA or later release.

In the PD client, a repository user has the List permission defined at the package level of a model. When logged into the Portal with this user and accessing the model via the Repository Explorer, the permissions assigned for the packaged are not applied. It appears that the permissions assigned to the model are cascaded down via the packages.

This occurs for any model type

To reproduce:

1) create a CDM

2) create a package

3) add three entities to the package

4) add all three entities as shortcuts on the Model default diagram

5) check in model

6) in the package, change the name of entity_3 to test_3

7) check in the model

NOTE: the change is not reflected on the entity_3 shortcut

add an attribute to test_3

9) check in the model

NOTE: entity_3 shortcut has been changed to test_3

NullPointer Exception when replicating SecureFile

LOB data.

an informative error message when initialization

fails due to required permissions not being set.

The customer does not like the fact that PD checks out (extracts) the model to the last saved location that is stored in the model. There attempts at defining a named path variable does not work for them.

incorrect error message when querying a LOB value

against the primary database fails.

a Null Pointer Exception when replication

is suspending.

same name but different owners, and if we have

already marked one of them for replication and

attempt to mark the other, the second mark

operation will fail.

definition by rs_drop_repdef,

the action is not put into the log.

table with more than one LOB column and if more

than 1MB of data is inserted into any of the LOB

columns, replication will fail.

error "Replication Server refused the following

invalid LTL command" when replicating a table

containing a NUMBER column and the value

contains an incorrect number of digits.

LOB processing.

the text/image column has different column name between primary and replicate.

- Repository on Microsoft SQL Server 2008

- PD OS: Windows XP SP3

This is for Itau-Unibanco

Some of the users are experiencing the following error (sporadically) when connecting to the repository for the 1st time:

PowerDesigner - Error:

Microsoft SQL Server Native Client 10.0

The object is in a zombie state. An object may enter a zombie state when either ITransaction::Commit or ITransaction::Abort is called, or when a storage object was created and not yet released.

The user is unable to connect at this point. Then, same user is able to connect at the second attempt with no error messages.

<scn>" after "suspend" and

when there is no archive log in the archive path.

automatic generated replication definitions when

an unmark command with force option is issued and

automatic replication definition setting is

enabled.

the LOB column replication status after

re-initialization with the 'force' option.

fails with a NullPointerException when there are

marked objects in the RASD that no longer exist

in the primary database.

Repository on Microsoft SQL Server 2008 x64 on Win2k8 x64

PD workstations on Windows XP SP3 machines

Issue description:

This is for Itau-Unibanco.

After migrating repository from 15.2 to 16.1, customer is experiencing slow response times from repository when checking in/out models that use customization.

However, customer confirmed checkin/checkout is very fast if they exclude their customization from within Resource Files folder.

Archivelog list in rman does not refresh

after log truncation with rman.

from 15.7.1 to 15.6, Replication Agent fails to

resume replication.

Repository on Microsoft SQL Server 2008 x64 (on Win2k8 x64)

PowerDesigner clients on Windows XP SP3

This is for Itau-Unibanco

After migrating repository from 15.2 to 16.1 EBF#8, the check in of some migrated models into same Folder it was before may result in the following error:

"There is already a document that has the same 'LDMD' in the folder Folder_3'

This error is random.

Should be possible to see Matrix Diagrams in the Portal "Repository Explorer" in the Diagram Tab if using "Save Diagrams for PowerDesigner Portal".

when replicating stored procedures.

when pdb_auto_create_repdefs is set to true,

erroneous messages indicating the replication

definition associated with the shadow table is

being dropped will be written to the log.

using the full replicate name; it is ignoring

the schema.

The value of lr_dump_marker

is not cleared after xlog

removed.

could be down due to Java IllegalStateException when

processing LOB records.

Description can be selected for Object Type 'Entity'.

user issues the "resume" instead of the

"resume resync" command.

(named, say, OWNER1.SEQ1) in the primary database

and marks the sequence in Replication Agent, and

then drop the sequence from the primary database,

executing 'pdb_setrepseq OWNER1.SEQ1' shows the

sequence is unmarked while executing

'pdb_setrepseq' shows it is marked. Then if the

user continues to create the same sequence and

mark it again, executing 'pdb_setrepseq' would

show two lines corresponding to OWNER1.SEQ1.

Replication Agent replicates

corrupt unicode characters of

LOB column if the primary Oracle

database version is 10g.

OS: Windows XP SP 3

Using MS Active Directory as LDAP system

Customer is facing the same exact repro steps as from CR#708924 which had been fixed in PD 16.1 EBF#7. By following same exact steps from CR#708924, customer is able to authenticate his LDAP user ONLY if leaving password as blank. Otherwise, he will run into same situation as described in CR#708924.

Repository on SQL Server 2008 x64

Customer would like the Integrate PD 15.2 feature to return on 16.1.

Starting in 16.1 local models are tight to a model in repository Information is saved inside the model at first check-in This is was made to ease model recognition when doing a check-in, and various action on models coming from different branches.

However, customer depends tightly on the Integrate architecture from 15.2 and this change affects his workflow.

Customer's text

If the models weren't checked in with the Save diagrams for PowerDesigner Portal, there is no way to set all models to have the diagrams saved for the portal without checking them out and checking them back in. This would be a nice feature to have.

OS: Windows XP, 7

Currently, PD allows LDAP authentication for Repository. Customer would like the ability of Windows Single Sign-on (SSO) to be implemented in product for future versions, so that same Windows login is able to enter repository without further authentication steps.

are not supported.

ct_results(): protocol specific layer: external error: There is a tds protocol error. An invalid numeric precision was received.

The maximum precision supported by Sybase Replication Server is 77.

To reproduce:

1) connect to a repository

2) select Repository->Administraion->Users menu

3) open properties of ADMIN user

4) in lower left corner, click on Property Sheet menu and select Impact and Lineage Analyis

NOTE: nothing happens

There appears to be no way to disassociate the new model.

To reproduce:

1) connect to repository

2) checkout a model

3) make changes to the model

4) select File->Save as New Model ... rename the file

5) check-in the model

NOTE: the model is still associated with the existing repository model

Creating Replication Agent

using resource file failed

if the version of primary

databases is Oracle 11gR1

or higher.

A command sent to the repository proxy failed.

07006 (SQLSTATE [Sybase][ODBC Driver][SQL Anywhere]Cannot convert '' to a timestamp)

Unexpected error while updating the repository document.

The check in of the document has been cancelled.

Portal shows different behaviour compared with PowerDesigner client GUI in the repository browser.

Due the login process appears the same whatever you decide to use SSL or not, this might lead to some confusion.

-> impossible to delete or to list properties

replicated through database subscription and some being replicated through repdef/subscription.

Repository on SQL Server 2008

Using Native SQL Server connection profile to connect to Repository (no proxy involved)

Issue description:

Users can磘 connect to repository any longer since yesterday.

According to customer, nothing was changed. However, PowerDesigner Portal and outside connections to SQL Server are ok.

We had customer turn on the profilers within PD machine and here what happens:

1) Customer chooses menu Repository > Connect

2) Customer chooses the SNC Native profile definition + Repository Login name + Password and click OK

3) At this moment the connection process transaction ends up by rolling back all below in DB trace profiler:

181 00:20:30.523 SSdbcStatement: Fetch 1 1

182 00:20:30.523 SSdbcStatement: Prepare: select RUSR,NAME,CODE,ISGR,SYST,RGHT,PSWD,CMMT,EMAD,OCUS,OCDT,OMUS,OMDT,PPRF,CPRF,SPRF,STTS,RNAM,LDAT,LFLD,NBFL,BLCK,UTMP from PMUSER order by RUSR 1 1

183 00:20:30.539 SSdbcStatement: Fetch 6.480 6.480

184 00:20:31.398 SSdbcStatement: Prepare: select NUMR,TLEN,TDAT from PMTEXT where TXID=? order by NUMR 1 1

185 00:20:31.398 SSdbcStatement: Fetch 531 531

186 00:20:31.867 SSdbcStatement: Prepare: select GRPE,RUSR,WGHT from PMGUSR order by RUSR,WGHT 1 1

187 00:20:31.913 SSdbcStatement: Fetch 65.185 65.185

188 00:20:34.538 SSdbcConnect: Rollbacking transaction 2 2

189 00:20:34.554 SSdbcConnect: SDBC Connection disconnected 1 1

190 00:20:34.554 SSdbcConnect: SDBC Connection deleted 1 1

191 00:20:34.570 SSdbcConnect: Rollbacking transaction 1 1

192 00:20:34.570 SSdbcConnect: SDBC Connection disconnected 1 1

193 00:20:34.570 SSdbcConnect: SDBC Connection deleted 1 1

194 00:20:34.585 SSdbcContext: SDBC Context deleted 1 1

4) Then connection screen loops to step 1

This is not a connection issue because "Test Connection" goes fine for his SNC native profile. Actually, the above transaction is rolled back and customer can never connect.

NullPointerException.

PD 16.1 EBF 11

PDV 16.1 EBF 4

Using both PD and PDV, I am able to connect to the same repository using the ADMIN login.

When attempting to connect using LDAP, the PD connection succeeds and the PDV connection attempt fails with the following message:

---------------------------

PowerDesigner - Error

---------------------------

Invalid LDAP user or password!

---------------------------

OK

---------------------------

If the user name happens to be the same as the CN name, the LDAP connection will work.

If samAccountName and CN are different, some users fail to connect.

Request is to be able like in the PD client to switch between different repositories.

and the extention is located under "My documents" -> "PowerDesigner Libraries" -> xyz Folder. This extention is delteted after the first check in.

It will inform user that he should not perform changes in these packages due he will not be able to update them into the repository due to not enough permissions.

It also applies to projects when a user has permissions on some models but not on some others.

The customer is using a cluster environment for their OID implementation. As a result of the cluster, a minimum of two calls to the OID server are required to validate the login & password. The first call obtains the correct server that will validate the user. The second call, would then pass the user credentials for validation.

E. 2012/12/18 19:01:24. ERROR #1028 DSI EXEC(106(1) UW2_standby.UW2) - dsiqmint.c(4445)

Message from server: Message: 102, State 181, Severity 15 -- 'Incorrect syntax near ...

This syntax error is result of an internally generated "writetext bulk owner.table.column" command where table/column names should be quoted but are not. This internally generated command will not show up in dsi_buf_dump.

require supplemental logging of unique indexes

and primary keys only for those tables that are

marked for replication and not for the entire

database.

Replication Agent pdb_execute_sql command.

ORACLE ONLY: Replication Agent downgrade accept(import)

process fail, when downgrade from RAX 15.7.1 to RAX 15.6.

did not work correctly

for Delete operation

workaround ... the workaround to the issue of using a derived class(db2_derived_class) with parent class of rs_db2_function_class instead of the default provided class of rs_db2_function_class is to query the rs_translation table this way ...

select classid from rs_classes where classname = 'db2_derived_class'

select classid from rs_classes where classname = 'rs_db2_function_class'

select * from rs_translation where classid = ( select classid from rs_classes where classname = 'rs_db2_function_class')

...... then create an insert for each row returned from the 3rd select above using all the same data except changing the classid column value to the value that comes back from the 1st select above ... shutdown and restart repserver ...

To reproduce:

1) Create a folder in the repository

2) add 3 or 4 branches under the folder

3) create two users - alpha (has manage all documents right) & beta (does not have the manage all document right)

4) connect to repository as alpha and attempt to delete a branch

NOTE; this action is successful

5) connect as beta and attempt to delete a branch

NOTE: the delete option is not available

One model cannot be updated into the repository anymore.

Error: "There is already a model with the same internal identifier in the folder ...".

IndexOutOfBoundsException.

Customer beleives that this is a product defect. However, this appears to be the default behavior but I have been unable to locate documentation supporting this.

any Oracle archive log files if the primary

Oracle is using ASM.

E. 2013/02/05 20:35:39. ERROR #1028 USER(sa) - seful\cm.c(8092) Message from server: Message: 30291, State 0, Severity 19 -- '[[Message Iteration=1|Data Source Name=sl6w2k8|SQL Function=CALL|SQLState=42000|Native Error=15002|Message=[Microsoft][ODBC SQL Server Driver][SQL Server]The procedure 'sys.sp_indexoption' cannot be executed within a transaction.

nvarchar data type column containing NULL,

empty value and empty space data into

HANA DB causes Replication Server DSI to

go down.

If you open the menu Repository -> Acitivit Log...

You will find the delete and insertrow buttons in the header.

It is not possible to delte or to insert rows in this window.

all the following versions have shown this problem.

Replication Server/15.7.1/P/NT (IX86)/Windows 2008 R2/1/OPT/Sun May 13 19:45:32 2012

Replication Server/15.7.1/P/NT (IX86)/Windows 2008 R2/1/DEBUG/Sun May 13 19:48:18 2012

RS/15.7.1/EBF 20408 ESD#1 rsebf1571/NT(IX86)/Win2008R2/1/DEBUG/Thu Sep 27 15:50:44 2012

RS/15.7.1/EBF 20681 ESD#2 rsebf1571/NT(IX86)/Win2008R2/1/DEBUG/Fri Nov 02 11:27:20 2012

background information on the problem

a certain warmstandby becomes out of sync ... resync is needed ...

when attempting to drop the conenction, the repserver disappears without any msg.

attempt to purge the queue still results in the repserver crashing.

No error messages are in RS errorlog. the repserver just disppears.

replication of update and delete operations on

Oracle NUMBER data may fail if the NUMBER value

is in a searchable column or is more than

36 digits in length.

The default CSI error handler is active. This implementation only intended to be used in non-production code. All consumers of CSI should override this implementation at factory creation time using the sybcsi_factory_set_error_callback() function.

Sybase CSI Warning Message: The Certicom FIPS runtime has already been initialized. The initialization will be skipped.

Agent log.

Failed to obtain a Thread for session

<com.sybase.jdbc4.tds.SrvSessionExt@4054a4>

Meanwhile, Replication Agent cannot be accessed via

isql client any more.

Customer don't like to have models with errors in the repository.

without restarting Replication Agent before the users

run 'ra_admin init' command, the command could fail

with the following error.

Command <ra_admin> failed - Could not initialize the

primary database because:

ORA-00942: table or view does not exist

Agent due to NullPointerException.

database is different from the character set of

Replication Server, and if you replicate a table with

CHAR or VARCHAR2 columns, Replication Server may be down

and fail to replicate the table.

We should get one only instance of the document in the browser.

This is a feature request to have these documents visible in the Workspace dropdown list box of the Check In Document window.

Customer would like a pop up screen that they can display messages on and which the user has to click "ok" to connect to the repository.

commit/rollback for transactions.

customer is seeing the following errors in their rs.log �

1)

E. ... ERROR #21 USER(foglight) - m/sqmprt.c(989)

Open server call to routine 'srv_xferdata' failed.

2)

I. ... No Sybase library information is available.

E. ... ERROR #21 USER( ) - xec/exec.c(457)

Open server call to routine 'srv_negotiate' failed.

3)

E. ... ERROR #60024 USER(foglight) - m/sqmprt.c(751)

Synchronization model MUTEX, element SQMD_Q_SEGS 115 1 was interrupted waiting for

a lock. (5295)eneric/useful/sync.c.

E. ... ERROR #60036 USER(foglight) - m/sqmprt.c(755)

Failed to acquire lock on sync element 'SQMD_Q_SEGS 115 1'. Info 7.

customer uses rs_monitoring frontend that connects to repserver and issues � admin who, sqm, admin who_is_down, admin health, admin disk_space �

feature request for better description of the situation in the repserver err.log ...

something like

a client application possibly abnormally disconnected from repserver before

repserver finished processing a request sent in by that vanished client application

and thus the errors are documention of the situation repserver is confronted with

because the client went away abnormally too quickly

feature request for better descriptions of these errs that are happening with

various different timings of the client leaving before repserver has finished

responding to it.

However, when user tries is trying to use the https certificate issued for the hostname of the machine (like it抯 supposed to be), the Portal object browser only shows repository folders and the models are missing in the treeview.

Replication Agent failed to be

initialzed due to the failure

of database trigger creation.

to return correct expired

date for graced license.

store procedure which has CHAR parameters, and the

replication definition is created by Replication

Agent through rs_create_repdef command, Replication

Server might be down during replication because of

the error 'ORA-23620: bind value size too large for

PL/SQL CALL operation'.

to resume replication after encountering

the error message similar to "Could not

set SCN generate ID to <65,536> for

locator <[GENID=0][SCN=12502638208][SCNGENID=0]

[THREAD=1][LSN=327298][BLK=337464][OFF=140]

[OATSCN=0][UNIQUE=0]> because the value exceeds

the maximum unsigned short value."

OS: Windows XP SP3

Customer is facing the same exact situation as described in previous CR#721850/CR#708924 which had been fixed in PD 16.1 EBF#7. By following same exact steps from CR#708924, customer is able to authenticate his LDAP user ONLY if leaving password as blank. Otherwise, he will run into same situation as described in CR#708924.

By running our internal utility to track down the issue, the utility authenticated fine. So, same routine needs to be ported to PD 16.1

parameters have different values in the configuration

files, if the user upgrades the Replication Agent

instance to a higher version, error might happen.

replicates an empty LOB column value as a

NULL value.

while Replication Agent is in replicating state.

operations that are logged just prior to an

Oracle online log switch.

the connection with Replication Agent,

Replication Agent may not be able to resume

replication without being restarted.

OdbcError(-2): [S1000] [Oracle][ODBC][Ora]ORA-00933: SQL command not properly ended

ORACLE ONLY: Replication Agent Oracle intervalDS

data type parse error make materialize session close.

The step to reproduce the issue:

1) Start Portal. Login

2) Click Search. Click New Search. In Model Type field, select Physical Data Model.

In Object type field, select Column. In Attributes field, select Data Type.

You can see Operator automatically set to Contains. There is no way to change it.

The customer wants to use '=' in this case.

execute SQL statements against the primary database

or RASD, in some cases, JDBC PreparedStatement or

Statement objects may not be closed to relinquish

database resources.

As an example, if you modify only a stereotyped metaclass, the Check in is cancelled because the repository document and the local copy are considered identical.

It does not happen with stereotypes.

processing.

data fails to replicate correctly in a double byte

environment.

location to a new location, Replication Agent

may go down when it is unable to find the

previous archive log file.

E. 2013/05/08 14:17:29. ERROR #1027 MatApply( ) - /dsiutil.c(420)

Open Client Client-Library error: Error: 34, Severity 5 -- '[ ExpressConnect for ODBC Data Source|statement ] /calm/bldr/rsconx/odbc/generic/src/odbccolumn.cpp:1241 odbcCol::SetValue failed, value too long. Setting Parameter 15 to it's bind location failed. '.

If an attempt is made to configure the cache size over the limit, Rep Server will crash during start up. This fix checks the calculated cache size and reverts to default values if it exceeds the limit and will then give warnings in RS log.

'ra_dumptran' when executed may prevent

other Replication Agent commands from

being executed.

configuration parameter lr_max_session_cache_size

to configure the session cache size.

lr_max_session_cache_size

The maximum number of Oracle sessions to be cached

in memory during replication.

Default:1024

Value: an integer from 128 to 2147483647

ArrayIndexOutOfBoundsException when replicating

large LOB operation from Oracle RAC node other

than the first one.

odbcStatement::PutData Data chunk of <x> bytes is larger than expected, total size exceeded.'.

This occurs because the character set conversion adds an additional byte per the encoding, which results in additional data beyond the initial data's length.

supports replicating 'truncate table'

command.

replicates a varchar single space value as an

empty string.

Implement feature to be able to configure more than one LDAP server for one Repository.

Step 5 states:

Extract the appropriate license library for your flavor of UNIX from

CMR_HOME/license/lib/sylapi 2.2.0.7.zip to CMR_HOME/license/lib."

However, if you install the Portal from PD 16.1 EBF 6 or PD 16.5, you must extract the file from the "sylapi 2.2.0.8fc.zip".

To reproduce:

1) Install Portal from 16.1 GA image

2) Apply portal EBF 6

3) Check for the referenced file, you find the following:

Under directory "C:\Program Files\Sybase\PowerDesigner Portal 16

\license\lib", there are two libraries files

"sylapi 2.2.0.8fc.zip" and "sylapi 2.2.0.7.zip"

NOTE: if you only install 16.1 EBF 6, the "sylapi 2.2.0.7.zip" is not installed.

due to incorrect quoted identifier handling by RepServer.

integrity guarantee for LOB data replication.

the replication definition correctly for a table

containing quoted identifier.

marked for replication in Replication Agent.

identifier, replication will fail when replicating

from Oracle to HANA DB.

primary schema to replicate schema mapping

when creating replication definitions.

replicating tables without a primary key.

and nvarchar data type column containing NULL,

empty value and empty space data into

HANA DB causes Replication Server DSI to

go down.

bulk loading operation causes Replication Agent to

go to REPLICATION DOWN state.

replicated incorrectly with the wrong

byte order.

Through the Portal UI, the customer cannot print a model that is larger than screen view or group of model objects that are larger than screen. Client requests the capability of printing models or parts of models that are larger than 8.5 by 11 thorugh the Portal

Parse schema file failed. [c/subcmp.c(8964)]

solution: upgrade to 64bit Replication Server 1571sp101 on linux

of primary key in the replication definition.

generates a fixed non-null LOB column as a null

LOB column. If the column on the replicate is

also a non-null, this will result in Replication

Server抯 DSI to go down with an error trying to

insert or update NULL into a non-null column.

Document added by "Add Document" should be Integrated into a brach where the document already exist if usign "Integrate Document". PD should open merge dialog.

Observed behaviour:

"Integrate Document" creates a second document with identical name in the a branch

Error: When client checked in a user profile, it is unable to edit in Repository tab.

If the primary table does not have a unique key and is using direct_load materialization, no activity is allowed on the primary until the subscription is VALID. If activity occurs during this process and 'dsi_check_unique_key' is 'on', the subscription will fail.

By turning off 'dsi_check_unique_key', the subscription may be created successfully but risks potential data inconsistency. Only turn off this configuration parameter under the following circumstances to ensure data is correct following subscription validation:

1. There cannot be any delete statements at the primary table with no unique key.

2. There cannot be any insert statements of duplicate rows at the primary table with no unique key.

all LTL operations other than BEGIN TRANSACTION

are null when ltl_origin_time_required is set to

true.

Replication Agent may fail to mark all

tables when configuration parameter

pdb_support_large_identifier is false.

The customer has been made aware that they can execute Check Model at any time via the Tools\Check Model menu.

Customer's business case:

We need to ensure that our metadata is of high quality. This includes the enforcement of consistent values in extended attributes. If the check model process could be invoked before model check-in, then the checks that we have put in-place, when they throw errors, will prevent bad quality metadata from being stored in the Repository. We are just now starting to audit the quality of metadata stored in the repository and find it quite poor at the present time. This situation needs to be improved.

to start on Windows x64 platform.

$ java -version

os::malloc: failed to malloc 1048576 bytes (errno 12).#

# There is insufficient memory for the Java Runtime Environment to continue.

# Native memory allocation (malloc) failed to allocate 1048576 bytes for /bas/sapjvm/cons/6/src/6/hotspot/src/share/vm/utilities/taskqueue.hpp:383: E/bas/sapjvm/cons/6/src/6/hotspot/src/share/vm/utilities/taskqueue.hpp. Out of swap space or heap resource limit exceeded (check with limits or ulimit)?

with resync option due to NullPointerException.

workaround

hard code RepServer to always return

varchar (3) 'yes' 0x79657

******************************************************************************************

use RSSD

go

rs_helpclassfstring rs_msss_function_class,rs_get_errormode

go

Note: the above rs_helpclassfstring should report the default "select @@ReturnNativeError"

update rs_systext set textval = "select 'yes'" where parentid = 0x000000000ffff825

select * from rs_systext where parentid in (select fstringid from rs_funcstrings where funcid = 0x0000000004000025 and classid = 0x0000000001000011 and fstringid = 0x000000000ffff825)

go

rs_helpclassfstring rs_msss_function_class,rs_get_errormode

go

Note: the above rs_helpclassfstring should report the new value for rs_get_errormode which is

select 'yes'

Then shutdown and restart repserver.

******************************************************************************************

E. 2013/08/05 09:49:49. ERROR #13045 MatApply( ) - seful/cm.c(4936)

Failed to connect to server 'HANARep' as user 'q06maint'. See CT-Lib and/or server error messages for more information.

E. 2013/08/05 09:49:49. ERROR #28153 SUB(sub_l_CUS_ACTOBT) - /sub_mat.c(1207)

Subscription <sub_l_CUS_ACTOBT> for replication definition <ra$UDBPrimary_Q04_SAPQ04_CUS_ACTOBT> with replicate at <HANARep.Q06> has experienced an unrecoverable error. The error flag for the subscription has been set

processing sends Replication Agent to

REPLICATION DOWN state.

the table in which a materialized view is stored

in the primary database to an ordinary table in

the replicate database.

processing sends Replication Agent to

REPLICATION DOWN state.

If database names are NOT set, the default database names 憁aster�, <SID>, and saptools will be used when setting up a new replication environment.

Once a replication environment has been setup its database list cannot be modified.

Added a new error action called "retry_forever":

- Any error with severity 17 will have this error action as default action

- It will retry forever

- It will sleep for DSI_RETRY_INTERVAL_C (14 seconds) between retries.

- It will print a message about "still retrying" every 2 minutes.

We will add a new error action called "retry_forever":

- Any error with severity 17 will have this error action as default action

- It will retry forever

- It will sleep for DSI_RETRY_INTERVAL_C (14 seconds) between retries.

- It will print a message about "still retrying" every 2 minutes.

different minor versions.

commit or rollback transaction notification to

Replication Server and that resulted in an open

transaction in Replication Server.

the pdb_archive_control_path configuration.

Workaround: Install RepAgent with a userid that has a $HOME directory with write permissions.

Workaround: Review Replication Agent documentation about upgrading the Microsoft SQL Server

database.

here is the error message where the upgrade process should be stopping ...

SENDMESG, spid ##: Msg 17001, Severity 16, State 1

Please complete your pending subscription create/drop before upgrading.

here is the the wrong error message where the upgrade process eventually stops ...

Message from server: Message: 208, State 1, Severity 16 -- 'rs_objects1500 not found.

Specify owner.objectname or use sp_help to check whether the object exists (sp_help

may produce lots of output).'.

parameter is configured with a tnsnames.ora

file which contains lowercase characters

in the TNS connection string, then when you

execute 'ra_config pds_host_name' or

'ra_config pds_port_number', errors may result.

work if Oracle archive directory is empty.

with version rao 156x the following lines used to be seen in the rao.log

I. 2013/08/30 16:52:21.067 INFORMATION com.sybase.ds.oracle.log.OracleRedoLogThreadR Oracle redo log thread reader for thread <1> is moving to log <68725>.

these msgs report progress rao is making reading through oracle's redo and archive log and are useful. add these same msgs in version 1571x.

previously marked for replication but has since

been unmarked.

hangs during bi-directional replication of

LOBs.

ra_set_autocorrection all, enable

may fail when there exists unmarked

tables in the primary database.

Replication Agent is unresponsive after several

invalid SQL statements are executed using

pdb_execute_sql.

unmarking a table.

table whose name contains special character(s).

Example: db_setreptable "a/special/table/name"

DOWN state with no article found error when the table

marking process is not completed successfully and

changes have been applied on the table.

incorrectly replicated as 0x.

RAO1571sp102 ra_locator gets stuck ... "ra_" row found in <prefix>_agent in pdb

'RA_' is the default prefix.

RAO1571sp102 ra_locator gets stuck ... "ra_" row found in RA_agent table in pdb (oracle)

formatting of a table that contains an

unsupported spatial type causes Replication

Agent to go to REPLICATION DOWN.

is different from the database name, XStream based

Replication Agent would fail to connect to the primary

database even if pds_service_name is correctly configured.

LogMiner based Replication Agent would fail to connect to

the primary database for same cause if its use_ssl

parameter is configured to true.

A change in the BEGIN_XACT log record body from

MSSQL 2008 to MSSQL 2012 can cause a parsing error

that stops replication.

Description:

Integrating from development Branches to a development Trunk should be simpler, suggest using the Change List mechanism to allow a streamlined process yet still retain control of the Trunk model.

Use of Change Lists in integrating branches to trunk

Scenario - customer has multiple branches with the same model. Users have permissions on their branch but not write permissions to the trunk. Use of change list should be implemented to allow branch users to submit an integration to be approved by the admin of the change list.

Enhancement to sync trunk changes with branches in the repository

Scenario - customer has multiple branches with the same model. When a user submits a change to his branch and then integrates it with the trunk (main) model, they would like these changes to replicate or sync down to all the branch models.

Customer's comments:

We need a way to get changes (in the root\main model) integrated back to branches without removing work in the branch � it抯 too complicated to pick changes out manually. The integration back should omit anything touched in the branch and overwrite anything not touched.

Add functionality on a Check in / Check-out before the Merge window to bypass the Merge window if there are no conflicts.

For the functionality, client is requesting code that would automatically check for no conflict and skip the merge screen altogether.

Please see customer's document for more context for this enhancement request.

with this message:

Select of subscription data failed at primary

database <primary.oracle> for subscription

<subscription_name> for replication definition

<repdef_name> with replicate at

<REPLICATE.HANADB>

pdb_enable_concurrent_mark for RAO.

pds_admin_connection_pool_size supported

to configure admin connection pool size.

New configuration parameter

pds_sql_connection_pool_size supported

to configure sql connection pool size.

ra_killconnection command extended to

support close all sql connections and

close free sql connections.

ORACLE ONLY(XStream): The NullPointerException will

be thrown out by Replication Agent when process the

DDL like 'ANALYZE TABLE xxx'.

Remove Certicom libraries reference from Replication

Agent scripts.

ORACLE ONLY(XStream): The wrong DDL statement

which missed the last ';' had been sent out

by Replication Agent for the DDL like

'create function'.

values that are null causes error in the TDS

stream.

exec,exec_trace_commands

sts,sts_print_dbproc

dist,dist_trace_commands

cm,cm_show_connect_disconnect

sqm,sqm_trace_writes

dsi,dsi_trace_writetext

dsi,dsi_cmd_dump

TEXT, NTEXT) with an empty LOB value

leads to a NULL value in the replication

target.

As a consequence, neither this LDAP user nor the approvers (ADMIN, etc.) do not receive any email notification.

Oracle(RAO)/XStream does not support replicating

stored procedures that include autonomous

transactions. XStream reverses the order of

execution when sending the autonomous transaction

and the stored procedure operations to RAO, thereby

creating the risk that RAO would apply the two

operations out of transactional order.

The issue was addressed by comparing the underlying IP addresses of aliases and differing FQNs. The code also supports recognizing and resolving multiple IP addresses from different network cards on the host.

The customer has defined 90 days for the inactive time period. This is the policy set for all end users. The problem is that they do not login in with ADMIN on a regular basis. When they do, the ADMIN login is blocked. Although they can create an emergency login, they would prefer the ADMIN account to be free of this property.

DDL replication as part of

Cepheus DB subscription enhancement

feature, which enables hetegeneous

DDL replication. The main target

database is HANA.

For any new check-in using a Proxy connection profile, the last char of these blocks has been replaced by a space char.

It works fine using a Direct connection.

For example, the first TXID NUMR ends up with

[...]ABCDEFGHIJKLMNOPQRSTUVWXYZ12

and NUMR2 starts with 4567890abcd[...]

whereas we should have [...]ABCDEFGHIJKLMNOPQRSTUVWXYZ123 and 4567890abcd...

the JDBC driver is the version of 11.1, after the

users execute Replication Agent command

'pdb_execute_sql' or after direct load

materialization is triggered, there might be error

reported in RepAgent log that 'pdb_execute_sql'

command fails to execute because of the following

error.

"while trying to read the field 'lastRowProcessed'

of a null object loaded from an array (which

itself was loaded from field

oracle.jdbc.driver.T4C8Oall.definesAccessors of an

object) with an index loaded from a constant"

If this error occurs, the users could substitute

an Oracle 11.2 JDBC driver for the Oracle 11.1

JDBC driver to address the issue.

compressed row parsing stops replication.

Replication Agent for Oracle runs

on Oracle 12c on the same functional

level as Oracle 11g.

Replication Agent for

Oracle XStream does not

support DDL replication

on Oracle 12.1.0.1 due

to Oracle bug.

REPLICATION_DOWN during scanning the IBM DB2 UDB

10.1 transaction log.

As a consequence, the checked out model is saved again in the old location.

Since SAP ERP applications can use Truncate Table operations, users of the SAP ERP system using RAM for replication will need to download a patch of the SAP Kernel that is aware of the sp_SybTruncateTable. Instructions for performing this download and install can be found in SAP Note 1972365 - "Retry on error 4711"

Other applications that wish to issue truncate table operations against tables being replicated by RAM will need to call the sp_SybTruncateTable stored procedure themselves. The interface for the stored procedure is:

sys.sp_SybTruncateTable @tabname varchar(128)

where "tabname" refers to the table that is to be truncated.

when trying to shut down on HP-UX.

It is a regression because this issue does not occur in PD 15.3 EBF17 and in 16.1 EBF17.

The defect can be reproduced in PD 16.5 SP03 build 4302.

It does not occur if 'Preserve modifications' option is not selected during

commit or rollback transaction notification to

Replication Server and that resulted in an open

transaction in Replication Server.

the following message is repeating in RS errorlog:

I. 2014/02/04 09:43:04. 'STFindRequiredStream: found inactive stream(pdb) runflag(0x07) state (2) strm-ref(8) csum-ref(3).'

Workaround: Execute the sap_status path command from the primary RMA.

see model before and after in bug model

<<

Unexpected exception

.src\rmgdtb.cpp(5060): long__stdcall SRmgDbAccessManager::ConsolidateUpdModel(long,struct ISRepObject*,long, int, int, long*, class SString*, class SString*,struct SRmgConsolidationProperties *)

A consolidation action is not selected by default.

>>

The Merge Changes to Repository Document window has detected different list of points.

Workaround: Always provide the database parameter when executing the command.

Workaround: To create a remote HADR topology, run the setup replication command with the distribution mode set to local, then use the sap_update_replication command to switch to remote distribution mode.

Msg 57, Level 12, State 0:

Server '<repserver name>', Line 0:

Type 0 is not a legal type.

The work around is to create subscription without the <num_selects> selects option, so sthat number of select threads is 1.

Workaround: None

Workaround: Review the state of the environment and ignore the Failing Exception entry in the sap_failover result set.

Workaround: If you delete the DD tables on the replicate database, you will need to shutdown and restart the RMA.

Workaround: None

Workaround: None

Microsoft SQL Server which prevents LOB columns

being automatically marked at the time their

table is marked. The automatic marking of LOB

columns should be the default behaviour in SP100.

There is a workaround.

Workaround: To correct the invalid 'ltl; mode, issue the sap_update command for both sites when upgrading the second site:

sap_update_replication synchronization_mode, <logical host name #1>, async

sap_update_replication synchronization_mode, <logical host name #2>, async

The list of parameters in Advanced tab shall be the same when searching for objects wither in the local workspace or in the repository.

Previous locked version is not unlocked with "unlock after check in" option after a new / modified version is checked in. Happen if "check out to file" is disabled.

Replication Agent fails to

process Oracle ALTER TABLE

SHRINK SPACE DDL statement.

The Replication Agent for Microsoft SQL Server now

supports display_only as a rs_create_repdef

option.

in the auto generated rep def

rs_repdef_schema_map settings.

A table is page compressed and has a

varbinary(max) column. If the length

of the inserted varbinary(max) is small

(around 9Kb), then RAM does not

correctly replicate the data.

When Replication Agent replicates

LOB_WRITE operations, it might go to

REPLICATION DOWN status, reporting a

ConcurrentModificationException error.

The default value is "*** RepositoryUserFullName - CurrentLongDateTime ***".

This is a feature request to be able to define and associate a template to the Check in/out comment.

This template could also use GTL variables ( %USER% - %DATETIME%, etc.) and regular expressions.

When lr_parallel_scan is enabled and

the command ra_statistics is executed,

a NullPointerException is encountered.

Initially, there was a problem in a remote topology where the maintenance user password was unknown to the Replication Server. To resolve this problem the RMA changed the password when the failover command was executed. Changing the ASE password requires sso role. Many customers will not allow a single login to have both sa and sso roles.

To address this issue, the RMA no longer changes the password. It uses new Replication Server commands to copy the existing password from the local Replication Server to the remote Replication Server.

LOB column may fail to be replicated.

The user may encounter errors in

RepServer during update operations

when HVAR is enabled and the source

table contains LOB columns that were

not modified in the update operation.

\\neo\cir\...\

Support repro environment:

PD 16.5 SP3 PL1 and PL2 pre-release; Repository installed on SQL Anywhere 12

Reproduced using proxy connection

Customer environment:

PowerDesigner 16.5 SP3 PL1 & Proxy 16.5 SP3 PL1

Repository installed on MS SQL Server:

Microsoft SQL Server 2005 - 9.00.4035.00 (X64) Nov 24 2008 16:17:31 Copyright (c) 1988-2005 Microsoft Corporation Enterprise Edition (64-bit) on Windows NT 5.2 (Build 3790: Service Pack 2)

All connections to the repository are via the proxy

NOTE: in the attached doc, I have reproduce bug 1. I am still attempting to repro bug 2.

Temporary protection for updates to repository documents allows any user to bypass the protection when using the proxy. The end result is the repository losing track of the model.

Repro steps

1. Connected to repository as ADMIN

2. Check-in model several times to with freeze version after check-in selected

3. Disconnect ADMIN user

4. On PC 1, User1 -> Check out model (Version 3 of a LDM was used for testing)

5. On PC2, User2 -> Check out the same model (same version also)

6. User1 -> Modify the model

a. User1 -> Check-in the model and keep the Merge Window open

7. User2 -> Modify the model.

a. User2 �> Check-in the model; Before Merge window popup would appear with message 揇ocument in Exclusive access on your workstation. Do you want to bypass this protection�

b. User2 -> Say 揧es� and move onto the merge window;

c. User2 -> Accept changes and continued check-in.

d. User2 -> Status should show as successfully checked in. And in the repository Model should show with new version (V4

8. User1 -> After User2 has checked in the changes, Accept Changes(Step 9) and check-in changes to the model the model;

9. User1 -> After Check-in, status should show successfully checked-in.

10. Now Check the Repository (refresh couple of times also) � Model is not in the Repository any more.

In HADR remote setting, ASE has a connection to the local SRS for replicating data into ASE, and a connection to the remote SRS for replicating data out to standby ASE.

SRS hide maintuser password feature will reset maintuser password periodically. If local SRS changes active ASE maintuser password, remote SRS won't have the updated maintuser password and therefore cannot access primary ASE to apply NC transactions.

Solution description:

SRS provides two new RCL for Dr Agent sync up remote SRS maintuser password.

1. sysadmin get_maintuser_pwd_enc, DS, DB

2. alter connection DS.DB set maintuser_pwd_enc to "encrypted_password"

The first RCL returns maintuser password in encrypted form for DS.DB (local SRS)

The second RCL change maintuser password in encrypted form for DS.DB (remote SRS)

Models and other documents that are in the Library folder in the Browser Local tab are not in the very latest version.

internal standby scanner does not work.

Repository: SQL Server 2008 (using native SQL Server driver to connect to SQL Server repository)

This is for Itau-Unibanco

Issue description:

------------------

Customer attempted to check in a pdm (BEL_MFT_SQL_Boleto Eletronico.pdm) into Repository in its latest vesion (version 130) and this version then disappeared in repository. According to customer, no errors in PD appeared.

into Microsoft SQL Server running in single user mode without having to

bring down other Microsoft SQL Server services.

Issue Description:

When UDB 9.7 with fix pack 5 or above is used, RAU will fail to load its native library. This would stop replication.

Cause Analysis:

This issue is caused by the IBM UDB 9.7 fix pack 5, in which UDB crypto library is upgraded and requires a recompilation of client applications that rely on it.

Details could be found at: http://www-01.ibm.com/support/docview.wss?uid=swg21590697

Work Around:

Shutdown RAU;

Request a new RAU native library from SAP/Sybase for your platform;

Overwrite the old library in directory $RAX\lib\$YourPlatform with the new binary;

Restart RAU;

. . .

T. 2014/04/01 18:34:01. (25): 'begin transaction '

E. 2014/04/01 18:34:01. ERROR #2056 DSI EXEC(106(1) rds16.rdb16) - /prs/sql.y(15109)

Line 1, character 21: Incorrect syntax with '='.

E. 2014/04/01 18:34:01. ERROR #5035 DSI EXEC(106(1) rds16.rdb16) - /dsiutil.c(2607)

Error from unpacker or parser. See previous message for more information.

T. 2014/04/01 18:34:01. (25): Disconnected from Server 'rds16', User 'rdb16_maint'.

. . .

Deleted LOB column in the primary

data server may not be successfully

replicated to the replicate data

server.

about the marked table when a user attempts to

mark a table that is already marked.

click "Show all authorized users" button,

then all users inherited permission will be shown "<None>".

And Effective permission is showing wrong, for example,

the user belonged to Administrator group, but effiective permission is still "Read".

Attached picture in bugmodel.

When replicating a large number of

rows that include binary columns,

Replication Agent may report an

OutOfMemoryError.

Replication Agent doesn't replicate

BFILE value when the table contains

SECURE FILE LOB columns in Oracle

11.2.0.3.

applying to table [nil][nil]...) to help to identify object

The versions tab of an repository object does not show the comments written during check in. It is only visible on model level. Would be good if the comment written during check in is visible as well on object level.

Replication Agent for Oracle fails to

filter Oracle system tables from the

pdb_find_tables result.

interface parallel scan performance was not

sufficient for customers with extremely high

volume.

The Replication Agent continues to scan the

database log but stops replicating data after an

ALTER INDEX command has been run against a

Replication Agent system table.

CSI uses OpenSSL for encryption

PD 16.5.3 and PD 16.5.2 includes CSI 2.8 M2 that is including a vulnerable version of OpenSSL

SQLAnywhere delivered with PD also includes OpenSSL

We must now deliver 1.0.1g or above

Rep Agent for Oracle goes into

Replication Down state when replicating

a delete operation on a table with a CLOB

column when that data has been loaded

using a charset that differs from the

database charset.

This only occurs on HP-UX with Oracle 11.1.06.

<<

Unexpected error while executing the following action:

Changing value of attribute Modification Date of object Target model reference

The check in of the document has been cancelled.

>>

There was a memory leak in SybFilter when the

SybFilter was running on a dynamic volume.

check it in

open it in web ui and do some changes : add a task, delete another one, and rename one

save it

extract the model from teh change list on PD desktop side

you got some diagram issues

* pool baackground color is blank

* message start event dos not have the corect icon

* the flow have stereotype displayed

* message format symbol not always good

symbols previously in the model are altered

new ones are OK

ECLogManager::svformat(): Error in formatting string.

Raised when tracing - ie, after:

> alter connection to DSNAME.dbname

> set trace to 'econn,dsi_buf_dump,on'

Repository User A:

1. Create a RQM and requirements

2. Create a BPM and a process

3. Link the process to one requirement

4. Check in both models to a repository

Repository User B:

1. Check out RQM from repository (read-only!)

2. Check out BPM from repository

3. Add two requirements to the BPM process

4. Check in the BPM

Repository User A:

1. Check out BPM

2. Verify process links in RQM requirements

-> Links are removed (see screenshots)

Replication Agent encountered the following error when replicating

LOBs:

"ERROR java.lang.IllegalStateException: Invalid type <Idle>

of LOB Cache Region."

partitioned tables if the Oracle recyclebin is

enabled.

and pds_tns_filename parameters for connecting

to Oracle, if the TNS Service_Name is different

than the Oracle database name, the Replication

Agent will raise a not initialized error.

Remove the requirement of pdb_archive_path for

replication, and just make the configuration

necessary for archive log truncation. Also

remove the property pdb_include_archives.

has a dependency on the Microsoft SQL Server

system resource database (mssqlsystemresource)

and thus no longer need to put

Microsoft SQL Server in single user mode when

configuring Replication Agent for replication

or when applying Microsoft SQL Server patches.

After upgrading RAM sp120 to sp20x, run ra_finalize_upgrade may fail due to ddl_password can't be decrypted correctly.

DDL statement is replicated and

applied in the target database

with a different user from who

executes the DDL in the primary

database. This happens in Oracle

to Oracle and Microsoft SQL Server

to Microsoft SQL Server replication.

It has been addressed for Oracle to

Oracle replication but not Microsoft

to Microsoft replication.

primary database will cause Replication Agent to

go into REPLICATION DOWN state with a

NullPointerException.

The Sybfilter driver causes windows to crash when it attempts to

add a log file path for an unmounted volume.

supplemental logging is enabled, and data is entered

into a table after a create table operation and there

is high replication latency, it is possible that

some rows may not be replicated.

implement ablility to search a string over all text attributes like (Comment, Description, Annotation etc.) over different kind and in combination of the version of an object.

for instance:

Search string like *<my_string>*

Sample: *MUREX*

The result should be:

PDM: xyz, and/or table: xyz, and/or column: xyz and/or

domain: xyz etc. and the version xyz (for instance the last version of the object)

archive logs until the oldest active transaction

in the database had been committed. When

transactions are open for several days

this causes a backup of archive logs on the

production system.

Example:

site2 Synchronization State Inconsistent Replication Synchronization Mode is currently operating as 'Inconsistent' instead of the configured value 'Synchronous'. Please refer to Replication Server logs for more details.

The synchronization mode is not being updated to the current state that Rep Agent is operating.

This does not affect the replication and the synchronization state is that reported by Rep Agent.

The work around is to re-start Rep Agent.

exec sp_dboption '<the name of the rssd>', 'select into', 'TRUE'

before a sp202 repserver is started the first time.

In addition, the primary user of the RSSD should have proper permission to add column rs_subscriptions.orderby_col.

Replication Agent fails to replicate any

transaction for a very large database.

If upgrading RAM sp120 to sp200/201, DDL replication will fail with error that complains system table object with id (e.g. 60, which is sysobjvalues) missing.

Open model test.bpm and edit diagram.

Sub-process has sub symbols. Edit Sub-process name in place,

input character 'proc' and select one term in auto complete list.

Then sub-process is replaced by shortcut. But its sub symbols are still there.

Save change. Go to workspace and open diagram in changelist.

But sub symbols has no name and no properties.

If parent object is deleted, should also delete sub-objects and their symbols.

When users are managed by LDAP, the PD documentation recommends creating user accounts in advanced. The customer is entering 100+ user accounts and then must open the properties of each user and then click the Check Name in order to validate the user.

The customer feels that PD should validate and update the user information by selecting the External option in the List of Users window.

But after import, boundary event is not at edge of its attached node symbol.

(see pic in bug directory)

that is already marked for replication, the LOB

column is not enabled for replication as expected.

Improve the performance for moving the truncation

point during high volume.

and if user has modified the path of the user-defined named path in the General Options on the client machine, the path may be reset to its default value defined in the General Preferences Profile after a few closes/restarts/re-connections.

PowerDesigner behavior does not seem to be consistent:

- either the value of a named path should be overwritten any time a user connect to the repository due to the check out of the .upf file,

- or the modification made locally should be always preserved.

Replication Agent doesn't distribute correct NULL

value for updating a BLOB column when Oracle

configuration parameter db_securefile is set to

'preferred' or 'always'.

- create and start 1 RS-ID RepServer (RSID)

- create and start 4 other RepServer (RRS1 to RRS4)

Creating the first 3 routes from the RS-ID server is fine

- create an outgoing route from RS-ID to 3 first replicate RepServer (RSID->RRS1, RSID->RRS2, RSID->RRS3)

When the fourth route is created from the RS-ID server a stack trace in the RS-ID server occurs.

- create a route from RS-ID to forth replicate RepServer (RSID->RRS4)

==>> RS-ID RepServer will end and stack trace.

and the table in this tablespace contains LOB

columns, Replication Agent does not replicate the

data applied on that table.

"ORA-01000: maximum open cursors exceeded" when

processing an UNSUPPORTED type of LogMiner record.

on a machine with resolution 1600x900

the dialog is too high

instance for Oracle using a resource file

with "pdb_include_archives=USE_DEFAULT"

configuration set causes start up to fail

due to an invalid value set for

"pdb_include_archives".

Specifically, they are requesting the following as a minimum:

1) Store and provide access to email notification history

2) Store and provide access to the contents of the emails sent by PD

3) Provide an option for the end user to modify the email templates

getting the Microsoft SQL Server

Log Sequence Number (LSN) on the end

of the log.

Improve the performance to start replication by

optimizing the method to get start of log.

Replication Agent when the volume of data that

needs to be replicated is a small part of overall

volume.

Any action in the Repository browser executes a query to check for changes:

PREPARE,select 1 from PMOBJT where OBJT=? and OMDT>?

This is a feature request not to send one request per package/object, but one unique request for

all objects of same class at once like the following one:

select o.code, o.name, o.omdt

from pmobjt o inner join pmclss c

on o.clss = c.clss

where o.clss = 1499

Open and edit diagram in portal.

There is a flow linking two tasks.

Change Task_1 to Sub-Process, change Task_2 to Call Activity.

Then try to select flow symbol, but it is difficult.

Even it is selected, the selection dotted line has different path from flow symbol

(see pic in bug directory).

Unable to allocate memory. [c/subcmp.c(4958)]

But in context pad of 'Normal Gateway', you can change it to 'Normal Gateway'.

Check out saved model in PD fat client (gateway.bpm in bug directory),

stereotype of 'Normal Gateway' is 'Stereotype'.

Corresponding symbol of exclusiveGateway 'Exclusive Gateway' has a property

'IsMarkerVisible=true', so it is imported as Exclusive Gateway.

Corresponding symbol of exclusiveGateway 'Normal Gateway' has no property 'IsMarkerVisible'

(which means its value is false), so it should be imported as Normal Gateway.

But it is imported as a gateway without stereotype.

Open and edit diagram in portal.

Draw a sequence flow from boundary event to task, and then save.

Check out saved model in PD fat client.

Double-click on that flow to open its property sheet,

but get following error, and flow has no source.

But this person never connected to PD repositroy and the I number is not declared

Send should not be grayed

When user click on send a message should warn him that the user does not exists

or no user is selected to send the invite or no invite text has been entered

1)display error

2)put focus in input text field , select text if any

Fixes for Sybfilter and Sybfilter Manager:

- Sybfilter causes Windows Server to crash

- Memory leak in Sybfilter when Sybfilter is

running on a dynamic volume.

- Sybfilter Manager 'check' command returns

ERROR: Unknown disk format error.

when user configures the pdb_archive_path

to an ASM path and does not specify the ASM

connection related properties.

But some flow symbols are not correctly imported, they only have middle pints

(see pic in bug directory).

This bpmn2 file is exported from NetWeaver, so only middle points of flow are

exported as waypoints of BPMNEdge, anchor points are not included.

BPMNEdge that have waypoints are not well imported.

Customer's text:

There is No transparency on who is using Check in and no notification on when it's done

We would like to see in the UI the USER who has a model locked (whilst performing a check-in/merge, for instance). There should also be a way to get a NOTIFICATION in the client when the model is UNLOCKED.

For example,

A LDM ha 50+ versions in the repository.

If you right click on an entity, it has 15 versions.

If you select version 10 of the entity, there is no way to determine which model 'version' this change was made.

PD should provide functionality that allows an user to right-click on version 10 of the entity and then identify (or allow to check out) the version of the model where this change was added to the repository.

Customer's text:

We need to be able to link changes for a repository object (e.g. an Entity, Attribute) to the version of the model that it was changed in.

I can't see how to do this in the UI, all I get is the 'Versions' tab on the repository object, and the 'Versions' tab on the Repository Model.

Without this its VERY hard to find out when something was changed and more importantly, WHY.

As a workaround, please provide a sample SQL query to identify the object version and the model version where it was created.

but get error in console and import cannot finish.

Both files contains flow symbol from boundary event.

If remove corresponding BPMNEdge in bpmn2 file, import can succeed.

ERROR com.sybase.ds.oracle.logmnr.OracleOperationPrnextOperation Transaction context already exists for begin transaction id <OPID 0x####.########.####:####.########.########.####>, Timestamp <YYYY-MM-DD HH:MM:SS.M>, XID <0x####.####.########>, Serial# <####>, Session# <###>, Session Info <login_username=*** client_info= OS_username=**** Machine_name=****** OS_terminal= OS_process_id=##### OS_program_name=******@****** (*** **-**)>, User Name <***>, Object ID <#>, REL_FILE# <#>, DATA_BLK# <#>, DATA_OBJD#<#>, Operation <START>, Op Code <#> Rollback <#> SSN <#> CSF <#> SQL Redo <set transaction read write;> >.

causes Replication Agent to go to REPLICATION DOWN state.

Customer's text:

It appears that PowerDesigner won't create the log folder and thus fail to create a defaultlogger.log file if the folder hasn't been created upfront.

We would like to see an enhacement request whereby PD automatically creates the specific folder if it doesn't already exist.

In addition, if the log folder doesn't exist, or if PD failes to write to the log for any other reason, a warning should be written to the Windows Application event log.

The justification for the above is that such simple configuration issues could prevent essential log information from being written making troubleshooting in case of production problems very difficult.

To reproduce:

1) Install PD Proxy

2) Specify a non-existant folder for the log

3) Start the PD Proxy service

NOTE : the log will not be created.

SMTP is defined in the repository. When a new user is created via the List of Users, an email with the password is sent out before the user's email is entered.

To reproduce:

1) Create User_1 (not an LDAP user) and do not enter an email address

2) Open the properties for User_1

3) In the Output window the following message is listed:

Password has been reset and an email notification message has been sent to User_1

The problem is that no email has been entered for User_1.

In support's testing, the email is sent to the 'sender's email specified in the SMTP configuration. However, the customer states that this is not happening at their site. The emails are being sent to random users.

We have been unable to identify a use case to send the email to random users.

REPLICATION_DOWN because of an

IndexOutOfBoundsException during compressed row

processing for a table that has more than 127

columns.

during materialization.

operation processing sends Replication Agent to

REPLICATION_DOWN state.

It does not happen if the "Check out to File" option is not selected.

Display option 'Automatic Link Routing' is not selected.

Open and edit diagram 'Vertical BPD' in portal.

Click 'Change Pool to Horizontal' in Tools menu.

Draw a new Pool on diagram, and the new pool is horizontal.

Save and submit the change, then pool in changelist is still horizontal.

Publish the changelist, then open diagram 'Vertical BPD - v2' again.

But pool comes back to a vertical one. (see Vertical - v2.bpm in bug directory)

save it, do compare,

in property of compare,

the description/annotation is still shown as rft format rather than html.

do compare,

in property of compare, the difference of the description/annotation value are wrong.

For example to clear all value of description, but in "New Value", it is still shown there.

Check out model in changelist, pool 'VIP customer' has only one default lane 'Default',

but the lane's stereotype is not set.

During import, if a process has corresponding participant, it will be imported as a pool.

But if the process has no lanes defined, a default one will be created for it.

In this case, lane's stereotype is not correctly set.

There are two pools and a message flow linking them.

Select both pools and the flow, then move.

But flow is not linking the pools any more. (see pic)

Should keep related position of pools and the flow.

When you attempt to check in a model, you are unable to select a specific folder.

When you try to select a Folder, PD appears to timeout and while attempting to

populate the folder list.

The repository is installed on MS SQL Server 2008. This problem is occurring with

either a direct or proxy connection. The proxy is installed on the same PC as the database.

on a newly created and initialized RAX instance,

RAX doesn't replicate DML commands on a newly

created table.

when the database maintenance user is not aliased to dbo or

granted sa_role. ("Database needs upgrade")

truncate xlog remotely through the local Oracle RMAN

utility.

Open and edit diagram in Internet Explorer.

Click 'Task', and edit any property in its property panel,

like add comment, or set Compensation to true..

Click 'CallActivity_2', in its property panel, there is an ATTO attribute 'Implementer'.

Click this link, but a confirmation dialog pops up asking whether leave or stay

(see pic in bug directory.)

But it should not.

2. Change the new created task type to Sub-Process or Call Activity.

Results:

Symbol disappear, and with console error,

15:31:46 ERROR ModelHandler - Error while executing command 'Change End Event Type':

TypeError: Value being assigned to SVGPoint.x is not a finite floating-point value.

All previous operations are rolled back.

...

These messages for BPMN2 validation rule are hard coded

a comment "// TODO i18n" is idnicating that is still to be done

in a BPM diagram, call the invite user dialog

you type I, you have all the users that starts by I

type 0. even if you have several users matching I0 the list becomes empty

2. Open model contents in explorer page, switch to contents facet.

3. Switch to other contents then switch back.

Results:

Contents facet failed to load packages and gave error in console.

2. Open model contents in explorer page, switch to Diagram facet.

3. Click Conversation diagram links to navigate.

Results:

Navigation failed.

Hyperlinks in Uses and Impacts facet also have this issue

1>edit bugmodel diagram "main", there is shortcut task "Task_internal" on the diagram,

then add a new task on the diagram, save it, error dialog pop up.

"The diagram can not be saved.

Unknown attribute BPMN2.Compensation

on Action UPDATE. "

1> connect to oracle repository;

2> update BPMN diagram, such as add some tasks and flows, save into changelist;

3> go to workspace; the thumbnail in workspace is not match with the updated diagram.

1>edit one diagram, which is with a shortcut task, add a new task on the diagram, save it,

in PD fat client side, check out the changelist model,

the symbol of the shortcut task is blank rather than blue color.

Task_1 and Task_2 are linked by a flow, and they both have symbol on Diagram_1 and Diagram_2.

Open Diagram_2 and edit in portal.

Delete Task_3, then save.

Check out model in changelist in PD fat client.

But in Diagram_2, Task_1 and Task_2 and flow between them are not in correct symbol format.

(see pic in bug directory)

1>open diagram, change a flow destination, save it,compare, in compare diagram,

focus on the flow, in property tab->changes,

new value is "[object Object]",which is no sense.

Open and edit diagram_package (under Package).

Draw pool or data store on it, and save.

But pool, lane and data store are created under package (BPMN Model_result.bpm).

They should be under model.

Following are steps:

Create a new diagram, then draw new nodes on it and save.

On portal, everything is normal, but in repository browser of PD desktop,

there is only one empty model, but if check out, diagram is normal.

So I added internal key, then in repository browser, change list is normal

Check out to local, diagram is normal

But in local browser, all new created object are under a hidden folder 慏B Web Service�

by user with write permission, open the changelist, edit it, save it,

now the v2 of the changelist is generated.

Go on to edit the diagram, and save it.

Then go back Workspace, not include the last change.

Tried to delete local cache diagram folder, and restart Tomcat,

the thumbnail/svg could be rebuilt correctly.

Failed to assert 'len >= 0 && len < 255+1' at 2287, 'eneric/prs/packbin.c'.

Causes crash in _pak_unpack_binary_cmdval.

- Choose the Tools menu

The Tools menu does not display, user cannot use the Help.

in compare diagram, only select Create objects in filter,

but delete objects is still show out.

Attached the action video in bugmodel.

processing of compressed tables sends Replication

Agent to the REPLICATION_DOWN state.

create a new model and try to import bpmn2 file from bug directory

there is message saying the import has been successful but the diagram is empty

The error should be fixed and users should not have to update the catalog tables manually.

the symbol position will be completely lost

Draw a task and change it to sub-process.

Draw at least two tasks or flow inside of it.

Change sub-process to task, but sub symbols are not deleted.

Then save. But get error dialog:

==============

The diagram cannot be saved.

Invalid object identifier null

on Action CREATE.Invalid object identifier D9513B49-F1AC-4FE2-93E6-DC5382C91C5D

Invalid object identifier null

on Action MOVE.

==============

This is a feature request to enhance this dialog and be able to change these values by the means of a checkbox or a dropdown instead of opening the Check Out Document dialog for each document one by one.

2.1.3.2 Reviewing a Change List

Procedure

5. Perform the following actions on a document by right-clicking it or using the toolbar:

- Check Out - Check the submitted version of the document out into your local workspace to review it in more detail (see Checking Documents Out of the Repository [page 327]).

- Compare - Compare the submitted version of the document in the change list with the mainline version in the repository (see Comparing Models [page 239]).

- Find in Open Change Lists - Open the List of Change Lists filtered to display only open change lists containing the document.

There is no icon for "Find Object in Open change list" in the Documents tab of the Change list property sheet.

There is a "Find Object in Open change list" item in the context menu of any document included in the change list.

Create a BPMN diagram

Create a pool

Create a task

Add a marker (comment bubble, ...)

Move the pool

==> The marker is not moved.

- Use galilei BPMN sample

- Create a subprocess

- Resize the height to a very small value

- Resize the height to the previous value

==> The text is aligned on top and the + icon is not on bottom (it could be gone or overlap the text on top)

error in sybase.log on flexlm server

9:24:42 (SYBASE) TCP_NODELAY NOT enabled

9:24:42 (SYBASE) UNSUPPORTED: "PD_SHELLW" (1103 64CB 122F 8C0B ) SYSTEM@PARD30119944A.emea.global.corp.sap

(License server system does not support this feature. (-18,327))

During security validation of SAP PowerDesigner 16.5 SP04,clickjacking vulnerabilty has been found.

Please refer the attached screenshot POC.

It is a violation as per https://wiki.wdf.sap.corp/wiki/display/PSSEC/SEC-264

It does not happen during initial check in but when the model has been modified and is being updated into the repository.

Customer's description:

Configurations are currently just an unordered list of documents.

An ordered list of documents is required when the user needs to control the exact

versions of each document are loaded into the client as dependencies.

example:

Extension at version 4

LDM Model (requiring Extension of Version 4) at version 3

PDM Model (requiring LDM at version 3)

If you don't load in them into the client in this order, you can potentially get an extension or an LDM at the wrong version.

This is particularly important for automation purposes

This is a feature request to be able to customize the email and include additional information to identify which repository it concerns.

fails to be initialized on Oracle 12c with ASM.

For example, user requires a list of URL strings that represent issue management tickets (JIRAs), so he can link issue tickets to change lists.

IDENTITY column caused Microsoft SQL Server

to compress the row values in a slightly

different order, which Replication Agent

was not handling.

pdb_archive_path is not configured, Replication Agent

goes down because it fails to reposition to the last

SCN where it stopped.

Replication Server/15.7.1/EBF 23096 SP203 rs1571sp203/Sun_svr4/OS 5.8/1/OPT64/Wed Aug 20 08:50:19 2014

Having configured a classic Warm/Standby with a Logical, an Active and a Standby connections.

When issuing an "alter logical connection", RepServer stack trace on signal 11:

T. 2014/10/16 19:34:49. (96): Thread USER(sa) infected with signal 11.

...

T. 2014/10/16 19:34:51. (96): Dumping memory trace.

T. 2014/10/16 19:34:51. (96): ***************STACK TRACE***************

T. 2014/10/16 19:34:51. (96): *****thread USER(sa)*******

T. 2014/10/16 19:34:51. (96): pc: 0x0000000000399f40 + 00000000000001d8, exc__appDumpStack ()

T. 2014/10/16 19:34:51. (96): pc: 0x000000000039e6d8 + 0000000000000434, exc_terminate ()

T. 2014/10/16 19:34:51. (96): pc: 0x00000000001ca0c8 + 00000000000007d0, sun_svr4_catch_signal ()

T. 2014/10/16 19:34:51. (96): pc: 0x00000000002606f0 + 00000000000004cc, _ddl_configure ()

T. 2014/10/16 19:34:51. (96): pc: 0x000000000023f238 + 0000000000000574, ddl_exec_cmd ()

T. 2014/10/16 19:34:51. (96): pc: 0x000000000023eda8 + 00000000000003d8, ddl_exec_cmd_wrapper ()

T. 2014/10/16 19:34:51. (96): pc: 0x00000000003b5c08 + 0000000000000c8c, _ex_process_cmds ()

T. 2014/10/16 19:34:51. (96): pc: 0x00000000003b0768 + 00000000000038d4, _ex_process_lang ()

T. 2014/10/16 19:34:51. (96): pc: 0x00000000003a4830 + 0000000000001d68, ex_lang_handler ()

T. 2014/10/16 19:34:51. (96): pc: 0x00000000006e4fb8 + 00000000000000dc, srv__exechandler ()

T. 2014/10/16 19:34:51. (96): pc: 0x00000000006e46c8 + 0000000000000320, srv__runclient ()

T. 2014/10/16 19:34:51. (96): pc: 0x00000000006ea4f8 + 0000000000000234, srv__start_function ()

T. 2014/10/16 19:34:51. (96): pc: 0x7e800000 + 1add6, _thr_slot_offset ()

T. 2014/10/16 19:34:51. (96): *****End of stack trace.****

T. 2014/10/16 19:34:51. (96): Internal error. Attempting to produce a core file.

=> This issue does not occurs using RepServer 15.7.1.sp202

and get fixes for vulnerability published on OpenSSL 1.0.1h included in CSI version CSI 2.9 M12 delviered with PD 16.5 SP04 PL1 and below

an ImageparseException error while processing

DDL on Microsoft SQL Server 2012.

Checked out model that was modified cannot be updated into the repository:

Some errors occur during the update of the repository model.

Unexpected error while updating the repository document.

The check in of the document has been cancelled.

It happens with either a direct connection or via the Proxy.

When Replication Agent remotely accesses Oracle

database through the local RMAN utility, the

connection couldn't be setup if environment

variable "ORACLE_HOME" is not set.

Replication Server/15.7.1/EBF 22393 SP200 repmain/Sun_svr4/OS 5.8/1/OPT64/Tue Feb 25 23:15:52 2014

In an environment where RepServer is configured to work alone (without routes) and having 9 classic Warm/standby connections (9 times logical+active+standby)

For some reason, RepServer's parameter "num_msgs" is still set to its default of 45568.

On work load, RepServer report error #1030 and #21 from function "srv_putmsgq()" and stack trace:

H. 2014/10/27 09:00:01. THREAD FATAL ERROR #1030 REP AGENT(DBSU1.colis) - /errhand.c(1168)

Open Server error: Error: 16016, State: 0, Severity 15 -- 'No free messages.'.

N. 2014/10/27 09:00:01. INTERNAL ERROR #21 REP AGENT(DBSU1.colis) - m/sqmext.c(4491)

Open server call to routine 'srv_putmsgq' failed.

I. 2014/10/27 09:00:01. No Sybase library information is available.

T. 2014/10/27 09:00:01. (26): Dumping memory trace.

T. 2014/10/27 09:00:01. (26): ***************STACK TRACE***************

T. 2014/10/27 09:00:01. (26): *****thread REP AGENT(DBSU1.colis)*******

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003a23e8 + 00000000000001d8, exc__appDumpStack ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003a6b80 + 0000000000000434, exc_terminate ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003a5140 + 0000000000000244, exc_reraise_exception ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003a5420 + 00000000000007a4, exc_raise_exception ()

T. 2014/10/27 09:00:01. (26): pc: 0x0000000000397aa0 + 0000000000000dbc, err_raise_syb_lib_my_maps ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003974a8 + 00000000000005cc, err_raise_syb_lib ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000000eec58 + 00000000000006e4, sqm_write_message ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003c6998 + 000000000000490c, _ex_lti_cmd ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003bd000 + 00000000000011cc, _ex_process_lti_cmd ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003b86f8 + 0000000000003d58, _ex_process_lang ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000003acbd8 + 0000000000001cbc, ex_lang_handler ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000006a4180 + 00000000000000dc, srv__exechandler ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000006a3890 + 0000000000000320, srv__runclient ()

T. 2014/10/27 09:00:01. (26): pc: 0x00000000006a96d0 + 0000000000000234, srv__start_function ()

T. 2014/10/27 09:00:01. (26): pc: 0xe0900000 + 1c4dd, _lwp_start ()

T. 2014/10/27 09:00:01. (26): *****End of stack trace.****

RepServer configuration:

optionname objid charvalue

------------------------------ ------------------ --------------------------------

num_msgqueues 0x0000000000000000 100

num_msgs 0x0000000000000000 45568

...

(note: in this case, "num_msgqueues" seems also undersized)

enable delayed replication occasionally.

if the pdb_auto_mark is true and DDL replication

is enabled in Replication Agent, Replication Agent

may replicate a Replication Agent system table

trigger DDL event that should not be replicated,

causing Replication Server to go down with errors

similar to "trigger RA_CREATETABLE_TRIG does

not exist".

When the source object of the shortcut is renamed, and when user compares the repository model and the local model, no difference has been detected.

PowerDesigner does not update the shortcut name unless the model which contains the external shortcut has been modified.

This is a feature request to update all changes into the repository, including shortcut name/code.

For example, parameter 'parallel_dist' doesn't exist

in the result.

Command "sap_configure_rs Rot, RS, <parameter name>, on"

returns error when the parameter value has been set

to 'on'.

"Standby LogMiner failed to start up due to

ORA-01292: no log file has been specified for

the current LogMiner session" error even if

the log file exists.

process rollback operations when configured

for parallel scan, resulting in duplicate

rows sent to Replication Server.

This issue may happen when user suspends capture and log transfer for a CI connection at the same time.

When this issue happens, Replication Server crashes due to double free.

- the repository folders (.Library folder, repository folders, and change lists folder) are all visible,

- the repository and change lists folders do not show any document anymore,

- the .Library folder still displays library documents but these cannot be expanded,

- models at the repository root are visible but cannot be expanded.

Deactivated users should not be allowed to create a new change list.

"PowerDesigner - Error: Invalid repository user or password!"

This is a feature request to display another meaningful message.

For example, "PowerDesigner - User has been deactivated. Login not permitted at this time."

It should allow administrators to:

1. block new logins to the repository via a configurable exception group at a scheduled time,

2. force users to disconnect displaying a pop up message asking them to save their work during grace period. It could also be scheduled, e.g. start disconnecting at 10am on Saturday and allow

users to postpone for up 1 hour then forcibly disconnect them.

All the above options should be configurable via an API/CLI.

with "Win64 not supported error"

uninstalling on Windows platform in

Replication Agent 15.7.1 SP202 and

SP203 release.

individual rollback operations on Oracle 11.1,

resulting in rollback operations not being sent to

Replication Server causing duplicate rows

sent to the target database.

The same applies to tables with no PK, since HANARAX will use all searchable columns as the PK instead.

"Table 'rasysparams' not found" error thrown in

log file

Non-last version objects under a last version parent: 9

select count(*) from PMRLSH R, PMOBJT S, PMOBJT T where R.PRNT=1 and R.SRCE=S.OBJT and R.TRGT=T.OBJT and S.CLSS not in (1, and S.LVRS=1 and T.LVRS<>1

Under the Repository Checkout Parameters, the Packages parameter is limited to 'models only'. Thus there is no way to check out a Package only once the model is added to a Project.

get changed in an Update operation, RAM

will send out wrong LOB status.

Workaround is to add a table level repdef.

ltl with many null

the SP120 noff1 branch. See CR 749999.

table and the column number of the table is the

multiple of 30, then replication is down.

(but reports mis-leading declaration of success)

when logical connection name and active connection name are the same

(as they would be if the warmstandby were configured from an existing

primary connection.

ra_helpdevice is incorrect if the device is

stored under ASM.

datatypes are returned for varchar(max), nvarchar(max), and varbinary(max)

datatypes.

Workaround: Ensure that the replication agents in your replication environment

are version 15.1 ESD #1.

with error: "ChainedOperationException: Could not

concatenate chained operation because: Error

processing record with supplemental data whose

column index duplicates other data columns".

Workaround: use direct connection to repository

parameter is set to a value that does not

contain an underscore, Replication Agent will

create system objects with delimited names

in the primary database during initialization.

This may cause Replication Agent api commands

to fail.

NULLPointerException when attempting to read

a redo log on an AIX raw device.

when the database is inactive with respect to

replicated changes, even after reading to the end

of a log file or when reading at the end of the

log.

unexpectedly in its attempt to log an

error message. This can occur when

specific internal resources are not

available during the process of logging

the error message.

'ra_config pdb_archive_remove, true' fails with

exception "Could not find Resource Bundle

containing index: LR_BAD_ARC_DIR_ACCESS_WRITE".

case-sensitive replicate table name correctly.

may suspend and not terminate. This can

occur if 'rs_init' fails to properly

spawn a Replication Server process.

during DDL replication with error "The object whose id is

XXX is not a valid object." after the Replication

Agent has been re-initialized.

the 'assign action' values should be set

to 'warn' rather than

'stop_replication'.

immediately follows the insertion of an empty LOB

in the same location, fails with an error in

Replication Server indicating "The data row that

the writetext command is associated with is

missing."

infinite loop when executing the 'resume

connection skip transaction' command

which may cause the complete consumption

of RSSD log space or ASE locks. This

can occur if the ASE and the Replication

Server are both configured with

multi-byte character sets.

Daylight Saving Time transition are incorrectly

adjusted by one hour.

to Replication 15.0 or higher, an

'assign action' value is incorrectly

created for ASE error #41880 rather than

#418801.

lr_ntext_byte_order is hidden, and can not be

set by user.

could fail yet still allow the

associated transaction to incorrectly

commit. This can occur when the failure

is caused by insufficient ASE procedure

cache.

Initial testing:

PDM file 22.1 MB

Proxy(TBONE server):

Consolidation: 36 minutes

Extraction: 14 minutes

No proxy(TBONE server):

Consolidation: 20 minutes

Extraction: 13 minutes

during Daylight Saving Time transition are

incorrectly adjusted by one hour.

Daylight Saving Time transition are

incorrectly adjusted by one hour.

Support ASE dump transaction command with

parameter "standby_access".

operations because there is no object information

in the LogRecord" occurred against an "ALTER

PROCEDURE ... COMPILE" command issued against a

procedure whose owner is in the Replication

Agent's owner filter list.

allowed by the license, the Replication

Server will incorrectly start rather

than terminate. This can occur only

when using an un-served license.

IllegalStateException if Replication

Agent is never resumed.

mark all user tables in the primary database.

*** glibc detected *** corrupted double-linked list: 0x08a0d908 ***

Theses errors occurs on RS126E6 when the standby side of replication is stopped when data is being replicated.

will generate a DSI 32058 error in Replication

Server with a length mismatch. The error is

similar to "The value given for

'<repdef name>.<encrypted column name>' cannot be

translated from datatype 'binary' to the required

datatype 'varbinary'. Value length is '33';

Maximum target length is '11'."

key violation at the replicate database because

an insert rolled back at the primary database is

incorrectly replicated.

time an SQM callback is executed.

Although this is a valid state for the

RSI internal list, the SQM callback is

unprepared for this state which can

result in a failure of the SQM.

"Replication Definition Not Found" Exception may

occur if a table is re-marked for replication

using the pdb_setreptable command. Table is marked

once with the "owner" keyword and once without the

"owner" keyword.

with error "The date time string is invalid and

can not be parsed correctly."

with error: "ChainedOperationException: Could not

concatenate chained operation because: Error

processing record with supplemental data whose

column index duplicates other data columns".

with a NullPointerException on a 'resume' command

when the owner of the primary database can not be

found in sys.sysusers view.

a table when a synonym is passed as a parameter

in session repository version locators prevents

truncation.

function replication if the primary

function name and the replicate function

name differ.

Two new Replication Agent commands:

"ra_deviceoffset", "ra_helpdeviceoffset"

ra_deviceoffset

Description

------------------------------------------------

Note This command is available only for Oracle.

------------------------------------------------

Changes the raw disk device read offset for a

log device recorded in the RASD.

Syntax

ra_deviceoffset device_id, device_offset

Parameters

device_id The device ID is the Oracle redo log

"Group number".

device_offset The offset of the raw device,

from which Replication Agent starts log scan.

Examples

ra_deviceoffset 1, 20

This command specifies the raw device read off-

set to the log device ID "1" as: 20.

Usage

- When ra_deviceoffset is invoked, Sybase Replica-

tion Agent records the specified raw device read

offset for the specified log device in the RASD.

- To get information about log devices stored in

the RASD, use the ra_helpdevice command.

- When ra_updatedevices is invoked, device offset

will not be reset to default value.

The default device read offset is zero.

- If ra_deviceoffset is invoked when the Sybase

Replication Agent instance is in Replicating

state, it returns an error.

- The ra_deviceoffset command is valid only

when the Sybase Replication Agent instance is

in Admin state.

See also

ra_helpdeviceoffset, ra_devicepath,

ra_helpdevice, ra_updatedevices

ra_helpdeviceoffset

Description

Note This command is available only for Oracle.

Returns device offset information about primary

database log devices from the RASD log device

repository.

Syntax

ra_helpdeviceoffset [device]

Parameters

device

The device ID of the primary database log device.

Examples

Example 1

ra_helpdeviceoffset

This command returns device offset information

about all primary database log devices recorded

in the log device repository.

Example 2

ra_helpdeviceoffset 1

This command returns information about the pri-

mary database log device ID "1" in the log

device repository

Usage

- The ra_helpdeviceoffset command returns the

following information for each primary database

log device recorded in the RASD:

Device ID - the log device ID defined by the

primary data server.

Note For Oracle, the ID is the value of the

Oracle Redo Log Group to which this file

belongs.

Database name - the name of the primary

database associated with the log device.

Device name - the logical name of the log

device defined by the primary data server.

Server device path - the path to a multi-

plexed version of the log device.

Disk device path - the path to the log

device (at the standby site).

Disk device offset - the offset from which

Sybase Replication Agent starts scan Oracle

redo log in the log device. The value of

device offset displayed is the value of

"raw_device_offset" property in the Rep-

lication agent configuration file plus the

offset value set by command "ra_deviceoffset".

Disk device status - the current status of

the server device path (ACCESSIBLE, NOT_VALID,

or OPEN).

- The disk device path is the current value

recorded in the RASD. Sybase Replication Agent

uses the disk device path recorded in its RASD

to find each log device.

- For each log device recorded in the RASD, you

can set or change the disk device path with the

ra_devicepath command and set or change the

device offset with the ra_deviceoffset command.

- If you do not specify a disk device path (using

ra_devicepath), the value recorded for the disk

device path is DEFAULT, and Sybase Replication

Agent uses the value recorded for the server

device path to find the log device.

- The disk device status is updated by the Log

Reader component each time you invoke the

ra_helpdeviceoffset command.

- If ra_helpdeviceoffset is invoked with no option,

it returns information for all log devices

recorded in the RASD log device repository.

- If ra_helpdeviceoffset is invoked with the device

option, it returns information only for the

specified log device.

- The ra_helpdevice command is valid when the

Sybase Replication Agent instance is in

either Admin or Replicating state.

See also

ra_deviceoffset, ra_helpdevice,

ra_updatedevices

when set to false, sends all before column images

regardless of columns identified in the

associated Replication Definition.

scanning stops replication.

synonym, no information was given back when you

try to query marked table through its synonym.

a SQLException error when invoking the ra_migrate

command.

file processing fails with the error

"Device <+XYZ> does not exist" when the parameter

"pdb_archive_path" is set to an ASM disk group

name.

procedure to add new arguments, an

"Unsupported SQL type 1111" error occurs.

with the Replication Server, (2) an MSA database

is also registered with the Replication Server

and (3) a replication definition is created for

a table in the warm standby database, then the

active connection to the warm standby database

cannot be dropped even after the replication

definition has been successfully dropped.

Server causes Replication Agent to go to ADMIN

state with an error similar to

"com.sybase.jdbc3.jdbc.SybSQLException: Line 1,

character 120: Incorrect syntax with the keyword

'_bf'" because a delete operation row has all

NULL data fields.

populates an RSSD table with the value 'set

computed_dml on'. The correct value is 'set

dml_on_computed on'.

Resume failure may happen when executing

'pdb_xlog move,force' before 'pdb_init'

and 'ra_init'.

In case of this failure, please issue

command 'pdb_init move_truncpt' to reset

truncation point and 'resume' MRA again.

VirtualFileNotFoundException error is thrown after

the primary database is restored.

to replicate after the following sequence of steps:

1. Create a stored procedure and mark it for

replication.

2. Unmark and drop the stored procedure

3. Re-create the procedure with the same name,

mark, execute it.

processing, the article's version in the RASD is

not correctly updated which may cause some

operations to be missed when Replication Agent

re-scans the log.

reports 'Segments Allocated' for the new partition as '1' if the stored procedure is invoked without parameters. If the stored procedure is invoked with the new partition name as the parameter, the 'Segments Allocated' field is correctly reported as '0'.

does not return the value for the

'dsi_alt_writetext' parameter.

incorrectly marked for replication.

Support migration to mra151.

can not be automatically created for a table with

a column containing 'sysname' datatype.

the ability to read redo logs that are mirrored

and also remain replicating if it cannot connect

to ASM when attempting to refresh log devices.

pdb_setreptable can not

recognize tables with long

names having '_' and '%'.

The length of table name cannot

exceed 255.

replicating due to an error. The error log

displays an error similar to:

"ChainedOperationException: Could not set header

to operation... because the chained operation

header.... is already set."

The error is caused by a failure to handle a

rollback operation applied to one log operation

of a chained operation.

java.lang.ArrayIndexOutOfBoundsException: Array index out of range: 0

Size: 6" occurred with primary Oracle multi-block

undo enabled.

Update consolidation adds the latest frozen version of the document in addition to the document version you added first as a configuration member.

pdb_date can not return

the correct month infomation

if OS regional is Chinese.

thrown when replicating nested UDD type.

to replicate commands whose log record "Flag Bits"

are set to value 3600.

rollback record to Replication Server that

incorrectly removes the begin transaction record,

causing Replication Server to fail with error:

"#9022 DIST Cannot deliver a command since no

transaction is open".

table, or to add or delete columns, a

NullPointerException occurs.

'Custom' and then decides to navigate backwards to choose a different

installation type, the component subset is not properly changed to match

the installation type.

Customer's description:

Currently, PD keeps track of the changes we make at the object level and automatically places check marks next to the updates that we have initiated when consolidating with the repository. However, if there are multiple changes to the same object, PD shows all the changes and we have to select which attribute changes that we have generated. We need PD to keep track of the changes down to the attribute level and automatically check only those changes at the attribute level that we have initiated. We realize that this can be done by manually modifying the check boxes at the attribute level; however, when you have a team of 8 people working on the same model and updating the same requirements, (modifications from a review perspective, modification from a consistency perspective and modifications from a content perspective), the number of modifcations is great and the time that it takes to step through each individual attribute is extensive. Fixing this would greatly reduce people clobbering eachother's work.

with the Linux 'libXp' package by default. If the Linux 'libXp' package

is missing, the RS 1510 GA and RMP 1510 GA 'setup' installation command

will fail and the following generic message is generated ...

The installer is unable to run in graphical mode. Try running

the installer with the -console or -silent flag.

If the 'setup' installation command is invoked with the following

parameters ...

setup -is:javaconsole -log '#' @ALL

... this additional message is generated ...

Install.product.install,

com.installshield.wizard.StandardWizardListener,

dbg,

java.lang.UnsatisfiedLinkError:

/mnt/cdrom/JVM/bin/libawt.so:

libXp.so.6:

cannot open shared object file:

No such file or directory

To resolve this issue, install the Linux 'libXp' package. This can be

done on Redhat Enterprise Linux with the following command ...

yum install libXp

ra_maintid command throws NullPointerException

after Replication Agent is initialized.

pdb_get_tables and pdb_get_procedures

can not recognize table/procedure

with long names having '_' and '%'.

The length of table/procedure name

cannot exceed 255.

correctly calculate the rollback point for a

"rollback to save point" operation. The failure

to rollback some operations causes an eventual

'duplicate key' violation at the replicate

database.

replicate owner when the replication definition

is created automatically and when it is created

for replicating Oracle to ASE.

A new property "rs_replicate_owner_required"

allows Replication Agent to always supply the

owner in the replicate table name when the

replication definition is generated.

rs_replicate_owner_required

Description:

Determines if the owner should always be supplied

in the replicate table name when the Replication

Agent generates replication definitions, so that

Replication Server can distinguish between

replicate tables of the same name with different

schema/owner, and different schema/owner than the

replicate maintenance user applying the command.

Default:

true

Values:

true � The owner will always be included in the

replicate table name of generated replication

definitions.

false � The owner will only be included in the

replicate table name of generated replication

definitions if the primary name is also owner

qualified.

Comments:

- When the rs_replicate_owner_required is set

to true, generated replication definitions will

always be created with the replicate owner

and table name in the "with replicate table

named" clause.

Example:

create replication definition ra$0xca61_MyTable

with primary at rao.ora10g

with primary table named 'MyTable'

with replicate table named 'MyOwner'.'MyTable'

(

"QUANTITY" rs_oracle_decimal,

"ORDER_NUMBER" varchar(200)

)

primary key( "QUANTITY","ORDER_NUMBER" )

searchable columns( "QUANTITY","ORDER_NUMBER" )

send standby all columns

replicate minimal columns

- When the rs_replicate_owner_required is set

to false, generated replication definitions will

be created with the replicate owner

and table name in the "with replicate table

named" clause only if the primary table is

owner qualified.

Example:

create replication definition ra$0xca61_MyTable

with primary at rao.ora10g

with primary table named 'MyTable'

with replicate table named 'MyTable'

(

"QUANTITY" rs_oracle_decimal,

"ORDER_NUMBER" varchar(200)

)

primary key( "QUANTITY","ORDER_NUMBER" )

searchable columns( "QUANTITY","ORDER_NUMBER" )

send standby all columns

replicate minimal columns

state with error "Error processing

record with supplemental data whose column index

duplicates other data columns."

did not reflect the correct product

catagory 'Mirror Replication Agent'

but always refer to 'Replication Agent'

Oracle 9i because the Replication Agent does not

correctly recognize the desired archive redo log.

The error is similar to "An Archive file was not

found. Reason: Archive log file for LSN <xxx> is

not found. File <archiveLog>.

new version. As a result, there are no fixed

bugs to list in this section.

the primary database can be inappropriately

replicated when a rollback occurs in the middle

of several concurrent transactions.

rs_oracle_datetime when a year value exceeds 4712.

convert an Oracle date value when the year

portion of the date value is larger than 4712.

Also, the installer for Replication Manager 15.1 does not display a confirmation message to create a new directory if the target installation directory does not exist.

with force option causes three Replication Agent

internal tables to be marked for replication and

associated replication definitions are created.

Unexpected replication definitions for rs_marker

and rs_dump are also created when

'rs_create_repdef all' command is issued.

RS using RS 15.0 or higher to a

replicate RS using RS 12.6 or earlier.

This can occur when the primary RS is

upgraded from RS 12.6 or earlier to RS

15.0 or later and the site version for

the primary RS is set to 1500.

the debug-log file shows following just before the proxy crashes

2008-06-03.15:24:30.000719 warning - SClient::OnAccept: too many clients

a DDL command that recompiles a procedure. The

error is similar to "Failed to handle the DDL

operations because there is no object

information in the LogRecord."

are not skipped because the pdb_skip_op command

interface does not take into account Oracle

SUBSCN field values.

Description

This command is available only for Oracle and

Microsoft SQL Server.

Returns, adds, or removes record identifiers from

a list of records to skip in processing. The

format of the identifier is database-specific.

To support RAC, log location and truncation is

done by System Change Number (SCN) instead of

Record Address (RBA). The skip operations feature

must be familiar with SCNs and thread numbers to

truncate the skip operations list and to skip the

correct operation across multiple threads.

Syntax

pdb_skip_op [ add , identifier ]

pdb_skip_op [ remove, { identifier | all } ]

Parameters

add

Adds a specified ID to the list of

identifiers of records to skip.

remove

Removes a specified ID from the list of

identifiers to skip.

identifier

Identifies the log record to skip. The syntax

allows for either an operation locator or a

database-specific identifier.

all

Allows you to remove all IDs in the

list of identifiers to skip.

Examples

Example 1

pdb_skip_op

This command with no parameters returns a list of

the identifiers for the records you want to skip.

Example 2

pdb_skip_op add, id

This command adds an ID to the list of identifiers

you want to skip.

Example 3

pdb_skip_op remove, id

To remove an ID to the list of identifiers you

want to skip.

Example 4

pdb_skip_op remove, all

This command removes all the IDs from the list of

identifiers you want to skip.

Usage

- The pdb_skip_op command allows you to skip

problem records, thereby avoid having to

re-initialize the Replication Agent.

- Skipped records are written to the system log

as a warning message.

- The pdb_skip_op command is valid when the

Replication Agent is in Admin state.

- The format of the generic identifier

is the 64 character LTM locator representing the

operation to skip.

- The format of the database-specific identifier:

- For Microsoft SQL Server, the identifier is

Log Sequence Number (LSN) and it has the

following form:

virtualfileseqno:pageid:operationsseqno

where:

- virtualfileseqno is the virtual

log sequence number.

- pageid is the page number.

- operationsseqno is the sequence

number of the operation.

You can get the LSN using the Microsoft

SQL Server DBCC command, or you can parse

it from the locator. For example:

- LSN in decimal format: �489:112:27�

- LSN in hexadecimal format:

�0x00000489:00000112:0027�

- For Oracle, the identifier contains, SCN

(System Change Number), redo log thread, and

Record Byte Address (RBA) and it has the

following form:

wrap.base.sub, thread, lsn.blknum.blkoffset

where:

- wrap is the SCN wrap number.

- base is the SCN base number.

- sub is the SCN sub index.

- thread is the redo log thread number

that the operation occurred on.

- lsn is the RBA log sequence number.

- blknum is the RBA block number.

- blkoffset is the RBA offset into the

block where this record resides.

The values must be specified in the

pdb_skip_op command together, enclosed in

quotes, each item separated by a period.

For example:

'0000.012345678.0001', '1',

'0012.0000444.0000123'

All values can be described as a hexadecimal

by prefixing the identifier with an �0x� as

follows:

'0x000c.00001bc.000007b'

or

'0x0000.00BC614E.000A'

See also

ra_helplocator, ra_locator

error similar to "Oracle log scanner error ...

log record <> was not correctly sorted - locator

is not greater than previous log record locator".

during resume when the maintenance user specified

in Replication Server is not a valid user in

primary database.

with an error similar to: "Could not concatenate

chained operation because: Error processing

record with supplemental data whose column index

duplicates other data columns".

enabled, the operation issued by the maintenance

user may not be filtered by the Replication Agent.

containing column names that are not quoted on the

"replicate if changed" clause.

LTL command exception when the connection

to the Replication Server is lost.