Known Issues

Learn about known issues and apply workarounds for Adaptive Server.

Note:

Sybase does not include System Problem Reports (SPR) and Closed Problem Reports (CPR) with Adaptive Server Enterprise. You can search the Web site for solved cases. Click Support > Services > Solved Cases .

Table 1. HP-specific issues
CR# Description
566820 The backupserver and auditinit utilities for the HP PA-RISC platform incorrectly list "XXXXXX" in their version strings instead of the actual EBF number for Adaptive Server.

Workaround: None

551972 Using Java in the database to create a file fails on HP-UX.

Due to known issue in HP-UX creating a file using the Java in the database feature fails unless the working directory permission mask is set to 0777 or equivalent.

Workaround: Set the mask on the working directory to 0777. (This is not the same as the physical directory permissions in the file system. The physical directory can have any permissions mask that is appropriate for that directory.) This setting is only for the sybpcidb configuration properties, where you can configure a working directory with a mask that allows or prevents certain actions from within the JVM. Apply the mask using either an add command or an update command. For example:
sp_jreconfig 'add', pca_jvm_work_dir, '/some/path(ugo=rwx)'
or
sp_jreconfig 'update', pca_jvm_work_dir, '/old/path', '/newpath(ugo=rwx)'
550351 Executing sp_companion on high availability system may stop responding if Java in database is enabled.
Workaround 1: If the server is running with multiple engines, take the engine that is running the JVM offline before running sp_companion. Use the monitoring tables to determine if Java was used on the server and to determine the JVM engine affinity.
1> select * from master..monPCISlots
2> go
If the status is:

  • "INIT", then Java was not used on the server.
  • "IN USE", then Java has been used on the server. The engine column shows the JVM engine affinity.
  • "IN USE" on engine 1, take the engine offline:
    1> sp_engine "offline", 1
    2> go
Workaround 2: Use these cluster commands if Workaround 1 fails, the server is configured for a single engine, or if the server stops responding when executing sp_companion:
cmhaltserv -v <pkg_name>   
This command kills <pkg_name>.sh, which stops package monitoring, but does not halt the package itself. Check ps output to ensure <pkg_name>.sh is not running. <pkg_name> usually corresponds to the companion server where Java was used, and where sp_companion is about to be executed. When the secondary companion is in failedover mode, the <pkg_name> refers to the primary companion.
cmhaltpkg <pkg_name>
cmmodpkg -e -n <same_node> <pkg_name>
cmrunpkg -n <same_node> <pkg_name>

Do not shutdown directly using the isql command. Doing so may cause unnecessary failovers.

546883 Disabling network I/O for Java is not effective on the HP platform.

Disabling network I/O for Java allows some network-related Java calls to be executed. This is with the dataserver binary, and not with the diagserver binary.

Workaround: If network-related operations are not permitted, do not allow Java classes to be loaded into databases that use network-related functionality.

Table 2. Adaptive Server issues
CR# Description
564420 Out-of-the-box memory configuration defaults to 80% of physical memory on the machine.

When you create a new Adaptive Server, the out-of-the-box configuration feature prompts you to allocate a total amount of memory for Adaptive Server, and a default value that is 80% of the system's physical memory. The server does not start if the amount of memory specified is not actually available for use by Adaptive Server.

Workaround: Determine the amount of memory you need, and confirm that this amount is available for Adaptive Server.

560437 Replication does not work if primary table lacks owner information.

When using SQL statement replication on a table subscription-based replication topology, a data markup language (DML) executed on a replicated table owned by the database owner (dbo) is not replicated as a statement if the replication definition defined for the primary table does not include the owner information. In such cases, the standard replication used instead.

For a database subscription-based replication topology, a DML is replicated according to the settings in the database replication definition.

If both a table and database replication definition/subscription exist on a table, and the owner information is not specified in the table replication definition, a DML is replicated according to the settings in the database replication definition.

Workaround: Mark the table as 'owner_on' at the source database, and create a replication definition for the primary table using the owner information.

551577 Opening a database may cause a scavenging process to stop responding.

When opening a database at the same time a process is scavenging the DBTABLE of that same database, the process may stop responding. You see modules dbt_grab(), dbt_clean(), metrics_proc_dump(), metrics_dump(), and dbt_get() in the stack of the blocked spid.

Workaround: Increase the number of open databases and reboot the server. The number of open databases should be configured to be at least the value returned by the query 'select count(*) from master..sysdatabases'.

544889 Potential performance degredation when CIS and Java in database are both enabled.

If the Java-in-the-database feature is enabled in Adaptive Server 15.0.3 or 15.0.3 ESD #1, all processes that access remote servers through Component Integration Services (CIS) are redirected to the engine that handles processes using Java. This can have a serious performance impact in a multi-engine system where most of the processes use CIS, as all of the processes are running on a single engine.

Workaround: If CIS and Java in the database feature are not used by the same client, the -T 9403 trace flag may improve performance by distributing CIS connections across all engines.

Note: Do not use the -T 9403 trace flag if you are unsure whether CIS and Java in database features are being used together by the same client. Doing so can result in a stack trace.
439404 Creating Job Scheduler schedules for Adaptive Server with Simplified Chinese.

When you attempt to create a Job Scheduler schedule using the Adaptive Server plug-in for an Adaptive Server localized for Simplified Chinese, you occasionally receive an error indicating that the varchar datatype cannot be converted to a date.

Workaround: Localize the Adaptive Server for Simplified Chinese before configuring Job Scheduler.

439406 Configuring Job Scheduler Self Management for Simplified Chinese

When you configure the Job Scheduler Self Management feature for an Adaptive Server localized for Simplified Chinese, you occasionally receive an error saying that a varchar datatype cannot be converted to a date. The Job Scheduler configuration process stops at this point.

Workaround: Do not configure the Self Management feature for Simplified Chinese.

416569 Updating file names requires the file name to be included in the where clause.

When updating the filename column of a proxy table mapped to a directory of files, Sybase recommends that you qualify the update to specific files by including the filename in the where clause. Some file systems may rename a file by deleting and re-creating a new directory entry; therefore, the same file name may be updated multiple times if the update is not restricted.

For example, this statement could cause multiple updates:

update t1 set filename=filename + 'old' where filetype
= 'REG'

Workaround: Add a clause such as 'and filename like '%.c'.

401543 Running allow backward scan does not work.

The allow backward scan configuration parameter does not work in Adaptive Server.

Workaround: If you are encountering an increase in deadlocks while executing backward scans, and separate problematic scans:
  1. Select the required rows into a temporary table in ascending order.

  2. Select from the temporary table in descending order.

400784 Starting Monitor Client GUI fails when the LDAP server is used.

When you start the Monitor Client GUI in an environment using LDAP instead of an interfaces file, the Monitor Client GUI cannot connect to any servers.

Workaround: Create an interfaces file with entries for the servers that are monitored and use this with the Monitor Client GUI.

382874 Creating RSSD database fails with earlier versions of Replication Server.

rs_init in Replication Server version 12.6 and earlier cannot create an RSSD database in Adaptive Server Enterprise version 15.0 or later.

Workaround: If you do not use an embedded RSSD (ERSSD), manually create the RSSD database before using rs_init to create a new Replication Server. See the Replication Server Administration Guide.


Created April 6, 2009. Send feedback on this help topic to Sybase Technical Publications: pubs@sybase.com