Targeted CR List for ASE 15.7 ESD#2 The purpose of this document is to help Sybase customers obtain a general idea of potential bug fixes in future Adaptive Server Enterprise (ASE) EBF releases. Please be sure to read the Disclaimer and General Notes below. Disclaimer: This document lists the targeted (not committed) fixed CR list for the release listed above. Please be aware that the purpose of this posting is solely to provide you, our customers with estimated release dates and targeted CR lists. Sybase does not commit to releasing EBFs on the specified dates or to including the CR fixes in the said EBF. While every effort will be made to meet the said targets, changes can occur at any time. It is also possible (although unlikely) that Sybase may decide not to release an EBF that was previously scheduled for release.
Bug ID Description
sqlloc doest not set column sortfile in syscharsets table.
run dbcc checktable(spt_jtext) report 7952 error.
Creating 2 new ASE w/out exiting the syconfig, when creating the second server, click "Network Connections", there will be an entry already exists.
DDLGen does not generate bindings of constraints (sp_bindmsg) created on tables to user defined message numbers.
invalid sort orders when set default language to Chinese using syconfig.
A new option is provided to turn ON/OFF FDP currency in MDA tables.
sp_locklogin cannot be executed from a database other than the master database after HA is setup.
Under dbcclog prompt when running some dbcclog command, pressing ^C will cause the dbcclog command to return to the sybmon prompt rather than dbcclog prompt. Then all the loaded transactions need to be loaded again.
The MDA tables monOpenObjectActivity, monOpenPartitionActivity and monProcessActivity are enhanced to include IOSize1Page, IOSize2Page, IOSize4Page and IOSize8Page columns to track large I/O counts per-spid/per-user process.
A new MDA table, monSpinlockActivity, is introduced to track spinlock activity.
After changing the cluster backend server password, many sybcluster commands cannot print messages that indicate the connection authorization issue.
In ASEPlugin, a new tab page "Localization" is added into server properties pages to allow the user to configure localization information for SDC server.
The error message 3405 is modified as "Recovering database ' < Database name > ' (dbid < Database ID > ).". The enhanced message includes the database ID as part of the error message.
In the showplan, when calling the sql function, the function name has garbage on it.
Syconfig.exe can not do upgrade on winx64 platform.
sqlupgrade fails to work if old version server to be upgraded is registered to LDAP server, it always tries to copy the interfaces file entry to the new directory and quits as failing to do so.
preupgrade does not raise an error if the master device is too small and will pass the test, then while booting up the higher version server to apply upgrade, ASE will raise 110272 error.
A 13988 error, "ALTER TABLE < Table Name > failed. You cannot change datatype of column < Column name > to < Target Data type > when it is an index key.", will occur if an attempt is made to change the datatype of a key column to BIT.
When marking a table with at least one LOB column for replication and for using a replication index while such an index already exists (for example, orphan index left over or index created first with sp_reptostandby then again with sp_setreptable), the index re-created is corrupt. The command 'dbcc tablealloc' reports error 2525. Run-time errors 605 and 1133 may also be reported.
New feature to provide user defined optimization goal functionality.
SELECT show_cached_text( < invalid value > ) INTO < table > can generate invalid rows and the following query could cause stacktrace.
When running under traceflag 566 to prevent the truncation of trailing blanks during char - > varchar conversion in worktables and during hashing operations, a 102 message, "Incorrect syntax near ','.", may be raised when running sp_sysmon with 'cache wizard'. The error is raised in sp_sysmon_dcachestats.
EN templates should be displayed when no other templates are available for JS in the language the client machine is using.
monOpenObjectActivity.RowsInserted is wrongly updated even when the INSERT command gets aborted.
In rare circumstances, a 12323 error may be raised when running REORG REBUILD on nonclustered index of a DATA_ONLY LOCKED table.
A 7928 error, "Index < idx_name > is not consistent; found < num1 > leaf rows but < num2 > data rows.", may be reported when running DBCC CHECKTABLE with bottom_up option on a non-empty dol table having one or more indexes.
Per page compaction for DOL with fixed length rows is enabled for REORG RECLAIM_SPACE
In Japanese locale, the "Configure PCI" window of syconfig is not localized.
The stored procedure name may not be displayed correctly in message, "Stored Procedure < stored proc name > uses deprecated builtin functions as", when running the preupgrade tool.
sp_helprotect truncates the objectname when it is greater than 30 characters.
When an invalid database dump is first loaded, a subsequent attempt to load a transaction dump will not report a useful error message to the end user.
Spinlock contention for Resource- > rproccache_spin is huge due to large memory allocation tuning.
ASE loops indefinitely with the message "NT operating system error 1056 in module 'nt386\source\demo\scm_nt.c' at line 1263: An instance of the service is already running."
DBCC LISTOAM on an invalid index ID may failed silently without any error message.
The recovery REDO pass at boot time or during LOAD DATABASE is unnecessarily slow in cases where there is one or more very long running transactions.
A new interface is provided that enables ALTER TABLE DROP column DDL to optionally execute by deferring the time-consuming data copy path.
The message "current process infected with 11" in the module 'bufdlink' together with a stack trace which includes the modules 'pg_allocate' and 'bufpindirty' may be reported in the error log when a table was in the process of being created and ASE has previously hit a 1204 error "ASE has run out of LOCKS. Re-run your command when there are fewer active users, etc.".
When "procedure cache size" configuration key value is increased, Procedure Cache Pool is not honoring the Engine Local Cache trace flags while allocating the extra pages to engine.
In rare circumstances, the message "Invalid column length: 12. Value must be between 0 and 10 for 'all-pages' row with minimum row length of 64" together with a stack trace which includes the modules 'collocate' and 'translate_varcols' may be reported on a 64 bits ASE on IBM AIX, when the datbase was loaded from a dump of a different platform.
The message "current process infected with 11" in the module 'proc_find_subscribable' together with a stack trace which includes the modules 'lt_sqlrep_exec_polic' and 's_mustrecompile' may be reported in the error log when a stored procedure is marked for replication and ASE is checking if the stored procedure needs recompilation.
A 820 error may be raised when executing BCP-IN with all INSERTSs that causes duplicate key errors during aggressive HOUSEKEEPER garbage collection.
In rare circumstances a 3327 error, "During redo, the page timestamp check found page < pageno > needs redo when the log record is before the first log record in phase 1 that needs redo...", may be reported when loading a database dump that was dumped on a relaxed-durability database (RDDB).
When a table is being replicated to a table that includes an additional IDENTITY, an error "AOP not found in getAopt" is raised.
A 2601 error, "Attempt to insert duplicate key row in object ...", may be raised when there are two sessions running the same MERGE statement concurrently against a table with unique index.
A private install is always increasing the version of the .cfg file on each cluster reboot even when there are NO changes to the configuration. The cluster in question is using a 4K page size and the problem occurs if a cache pool of 32K is configured.
Enabling replication in a database can reduce the transaction throughput and can cause a high CPU usage.
The ASE PCI layer responsible for the execution handshake of ADT and UDF Java operations with the integrated COTS JavaVM is enhanced to achieve better performance.
The message, "Invalid column length:... . Value must be between ... and ... at offset ... for 'all-pages' row with minimum length of ...", together with a stack trace which includes the modules 'LeSubst', 'LeScanOp::_LeOpNext' and 'execproc' may be reported in the error log when the 299 command line traceflag is used and under some circumstances a stored procedure which queries a user temporary table created outside is executed with different schemas for that table but the same plan is reused.
In rare circumstances, the error 8203 followed by a stack trace that includes the module des_create_inxact() could be raised when rolling back a transaction.
A 803 error may be raised together with some diagnostics information if recovery fault isolation is configured to page level and more than 12 pages have been isolated and added to the suspect page list.
Under certain circumstances, running multiple instances of the third party monitoring tool asemon can lead to stack traces due to signal 11. The more instances that are running, the higher the chance of a re-occurrence of this issue.
Error 3606 (Arithmetic overflow) occurs if configure more than 2G procedure cache size by sp_configure "procedure cache size",0,"2G", instead of specifying the number of 2k pages.
A shared memory dump hangs if the ASE server encounters a signal 11 in the malloc system call.
If there are some NULL values in a TEXT column of a table, ASE may return " " instead of NULL when converting values in the TEXT column to CHAR data type.
The IOR_INTER_CALL error message is incorrect and is missing the device name.
The threshold status passed to the threshold stored procedure sp_thresholdaction may contain some unexpected data for thresholds other than the last chance threshold.
Any uppercase letter 'Z' in column name will be mistakenly converted into lowercase letter 'z' in the output of SELECT FOR XML.
When running DBCC PGLINKAGE with the target page being identical to the starting page, the target page will not be found and thus leads to a searching failure.
ASE may not release memory allocated to MEMC_MEMAUXBUF_1 in some circumstances such as when we use BCP to load data into table with large number of partitions. This may lead to error 701.
Dynamic Partition Elimination is not used for a store procedure with EXEC IMMEDIATE.
A procedure executing a function bound using des_bind encounters a 8242 error if a DDL statement is executed concurrently.
On Windows 64bit platform, if multiple network interfaces are available on the system, the Job Scheduler (JS) Task may fail to connect to JSAgent leading to failure of JS configuration.
Queries in parallel mode using trace flag 16953 may generate bad plans because costing for parallel operators is incorrect. This is due to incorrect costing of one of the cost components in the parallel cost object.
The message, "WARNING: The transaction log for database ' < dbname > ' (dbid < dbid > ) can no longer be dumped due to user ' < username > ' (suid < suid > ) executing < cmdname > ", should not be raised for databases with "trunc log on chkpt" enabled.
A new session level trace flag 11908 is added, if it's turned on, zero used/unused OAM entries will not be removed when running reorg compact.
The message "current process infected with 11" in the module 'upsleepgeneric' together with a stack trace which includes the modules 'ha_checkalive' and 'ha_vcs_getgroup' may be reported in the error log during the installation of ASE HA Cluster companionship on AIX.
Error msg 1579 is raised if the second char_expression is longer than 255 in builtin function compare().
New columns are added to the monCachedStatement MDA table to provide data on rows read, lock usage and sort activity.
Some performance improvements are made by 1) reducing client-server communication overhead when client passes parameters to dynamic SQL statements and 2) by optimizing the ASE code path for processing dynamic SQL batch commands when configuration option 'streamlined dynamic SQL' is enabled.
Recovery may fail with an 8201 error when the configuration parameter "per object statistics active" is enabled.
New feature to allow DUMP TRANSACTION to run concurrently with DUMP DATABASE.
The output of the LogicalReads column in the monProcessActivity table is always 0.
There are PCIDB spelling errors within the stored procedure arguments.
In rare circumstances on a heavily-loaded multi-engine system, a 8201 error "Keep count of descriptor (objid= < id > , dbid= < id > ) was expected to be 1. Instead 0 was found." may be reported in the error log together with a stack trace which includes the modules 'des__unkeep' and 'close_range' when an INSERT, DELETE or UPDATE is executed on a table that has many referential integrity constraints defined and after the execution of one of its triggers. Alternatively the message "infected with 11" in the modules 'seq_schema_check_virtab' or 'LeScanOp::_LeOpOpen' may be reported.
Additional syntax for the UPDATE STATISTICS command is implemented to support an ordered tuple mode for non-indexed columns. The new syntax is: update statistics (col1), (col2) .... (coln)
The Assertion ((varNo == VR_NOENTRY) || (addIn.varNo == varNo)) may fail for queries with parallel reformatting strategies where the reformatted table is partitioned N-ways.
If a new instance is dynamically added to the cluster in SDC mode, it may not be able to join the cluster when no other activity is taking place in the cluster.
The MDA table monCachedStatement is enhanced to include new AdjustToParallel, AdjustToSerial and ThreadDeficit columns to track parallel runtime adjustment.
The help text for system stored procedure sp_replication_path includes an incorrect syntax for the unbind action with a path as a target.
A query containing an OR clause nested under another OR clause might generate a SEGV. This could occur if the in_list_general_or optcriteria is on. The optcriteria is enabled in the ase1503esd4 optlevel and above. The stack will show that we're code generating (CgpGlobal::CgpGenerate()) and trying to resolve a relop (ResRelOp) for a predicate bitmap (CgpPop::_CgpPePredBmToScalar).
When marking a database with sp_reptostandby and 'use_index', some TEXT columns may be handled unexpectedly with a replication status do_not_replicate instead of replicate_if_changed.
UPDATE on DOL table with non-clustered index on two fixed length columns may fail with a 644 error.
Under rare race conditions, when BCP IN is in progress with the INSERTs running concurrently on the same table, BCP may encounter an 820 error where the process tries to dirty a buffer which is already pinned to another PLC.
ASE will fail to dump shared memory when 'number of dump threads' are configured to 2 or more. The errorlog will contain an infected with 11 stacktrace that includes shm_mergedump.
ASE running on Windows platform crashes on startup when the configuration parameter 'allow sql server asynch i/o' is set to the non-default value of 0.
High rprocmgr spinlock contention may be observed under heavy workload of dynamic SQL and stored procedures.
The message "current process infected with 11" in the module 'rec_undo_session' together with a stack trace which includes the modules 'xact__rollback_local' and 'crt_main' may be reported in the error log when the 1105 error was previously reported during the creation of a table.
When there is a declare cursor on deleted/inserted table in an instead of trigger on a view, INSERT/UPDATE/DELETE on those views triggers the error message.
When tempdb_space resource limit is reached, any subsequent work table creation operation fails and does not release the buffers and the allocated tempdb pages during the execution.
LDAP errors, such as "The LDAP URL string '' is invalid. LDAP subsystem reported error '(URL row not found in sysattributes)'", may be reported in the ASE errorlog at startup if traceflag 3635 is enabled. This occurs even if "enable ldap user auth" is 0.
The system stored procedure sp_reptostandby now reports the progress of its execution.
If the character set is UTF8 and a stored procedure is created with upper() or lower() builtin function and it has a LIKE clause to compare this upper() or lower() with another parameter, a 2431 warning message may be raised when the stored procedure is recompiled.
In ASE Cluster Edition, dropping a database and recreating a new local temporary database reusing the same database ID may result in issues during runtime.
In rare circumstances on a heavily-loaded multi-engine system, the message "timeslice < value > , current process infected" together with a stack trace which includes the modules 'jldr_ReleaseLocalLoader' and 'pci_free' may be reported in the error log when applications use JAVA extensively.
A 11060 error, "This statement has exceeded the maximum number of recompilations (10). This indicates an internal error. Please note the error prior to this one and contact Sybase Technical Support.", may be reported under some circumstances when a stored procedure containing a SELECT INTO ... UNION statement is executed and references session temporary tables and/or stored procedure parameters.
The stored procedure, sp_reptostandby, will now print a message warning that the execution can take a long time.
Merge partition into an existing partition but not a different segment is not supported, but the command will succeed even though the command takes no effect.
Trailing zeroes from varbinary (or binary null) columns in proxy tables are truncated even though 'disable varbinary truncation' is set.
In rare circumstance, the message, "current process infected with signal 11" or error 632, "Attempt to move memory with an incorrect length of < length > . Maximum allowed length is < max length > ." may be reported with stack trace after upgrading to ASE 15.7.
Under a rare situation, if a CONVERT node is needed for the like clause, it could return wrong result.
A 204 error followed by a stack trace involving the modules 'norm3_propagate', 'norm3__driver' and 'normalize' may be reported in the errorlog when a parameter, other than the allowed integer datatype, is specified for the identity() built-in.
In some circumstances in a cluster environment ALTER DATABASE can get a self deadlock.
An error 2836, "The cursor 'cur' is already tracking another procedure. This is an internal error.", followed by a stacktrace involving the modules 'curs_release', 'curs_close' and 'css_deallocate' may be reported when a cursor that needs to recompile is opened.
When using the distribution by connection model and "ddl path for unbound objects" is set to "default", DDL is not replicated to the correct path.
Job Scheduler does not print error messages in the error log without trace flag 3641.
When the proxy table contains a LOB column, executing an INSERT-SELECT queries joining a local and proxy column may hang or result in Ct-library errors such as "Ct-library: Msg 0x0101013e Fatal (Layer=1 Origin=1 Severity=1 Number=62): ct_get_data(): user api layer: external error:".
The SsqlId column in the monSysStatement MDA table incorrectly shows zero instead of the correct statement cache ID.
Error 11060 may be raised by an INSERT/DELETE/UPDATE into a view with an INSTEAD OF TRIGGER when SET CIPHERTEXT is ON and no encrypted columns on the underlying table.
In ASE Cluster Edition when the 'print deadlock information' configuration option is set or traceflag 1204 is turned on, for locks involved in the deadlock, incorrect identity may be printed when the lock is a Globally Unique Identifier (GUID) lock.
In extremely rare circumstances under Multiple Path Replication, it is possible that the retrieval order of path information from sysattributes could be different between Rep Agent runs causing log data to be sent to different Rep Servers which could lead to undetected duplicate records on the replicate database.
The message, " Expected object passed to kmuxtskGetThreadpool is not of type Multiplexed Task", is reported in the errorlog.
When ASE is configured with multi-bytes charset with 'statement cache' and 'literal autoparam' enabled, repeatedly executing a SQL batch with comments containing multiple byte characters can cause error 102.
Under some rare conditions, the message, "current process infected with signal 11", together with a stack trace involving the functions s__setup_tabsdes() and s_setuptables() or a 8201 error message may be reported when multiple processes are running stored procedures involving many tables.
The message "current process infected with 11" in the module 'LeRun' together with a stack trace which includes the modules 'LeBulkOpOpen' and 'bulk_main' may be reported in the error log when running a bulk copy operation into a table defined with materialized computed columns involving expressions using system functions like CURRENT_BIGDATETIME(), GETDATE(), etc.
Error 9275, "Could not locate the before image for the INSERT log record at ( < page_number > , < row_number > ).", may be reported by the RepAgent thread when a table that has compressed rows is marked for replication, resulting in the RepAgent thread terminating itself.
The 420 message, "The TEXT, IMAGE and UNITEXT datatypes cannot be used in an ORDER BY clause or in the select list of a query in a UNION statement.", may be reported when selecting a TEXT column from a proxy table in a UNION ALL query.
For the schedules running over midnight, Job Scheduler sometimes misses execution of first valid schedule job and sometimes executes scheduled job on invalid schedules time.
ASE running on Windows platform may exit with no message, stack trace or other indication in the error log when stack limits are exceeded.
The values of the ExclusiveLockWaitTime, SharedLockWaitTime and UpdateLockWaitTime columns in the monOpenObjectActivity MDA table are not updated when queries are blocked by locks.
In-place upgrade from older versions of ASE to ASE 15.7 may fail because the new ASE fails to start with error "The configured value '4096' for parameter 'kernel resource memory' is too low.". This occurs when ASE is configured with large "number of user connections" and "open indexes".
When executing the stored procedure, sp_reptostandby, the message warning that the execution can take a long time is not returned to the user immediately.
When the Job Scheduler is setup on a server whose master database is replicated, the replication of the DDL command executed on the primary ASE by the JS task to set the password of the 'jstask' user fails due to missing credentials in the LTL sent by RepAgent to Replication Server.
The DROP THREAD POOL command may lead to multiple stack traces, signals and timeslice errors
The Backup Server error, "The 'write' call failed for database/archive device while working on stripe device ' < archive name > ' with error number 22", may be raised if the trace flag 256 is used to enable DIRECT_IO on archive files and the dump is compressed.
The signal 11 together with a stack trace which includes the modules 'inserttext' may be reported in the error log when a LOB column with NULL value is updated to itself with replication mode.
In rare circumstances, a 644 error, "Index row entry for data row id ( < pg > , < row > ) is missing from index page < pg > of index id of table 'sysstatistics' in database ' < db > '", may be reported during the load of an 12.5 database in a 15.7 dataserver.
The message, "current process infected with 11", together with a stack trace which includes the module 'ulreleasespinlock' may be reported in the errorlog when the remote server is shut down abruptly using 'shutdown with nowait' when a transactional query is in progress.
Login account locked due to inactivity remains locked even after being unlocked by sp_locklogin.
In a 2 node or greater ASE Cluster Edition setup, a database with a log segment threshold defined may result in incorrect values for unreserved data pages and no thresholds enabled. The database will report log is full and transactions will either suspend or abort, depending on the configuration paremeter "abort tran on log full" value. This can only occur when the database has multiple thresholds and the second threshold is crossed before the first one is completely handled.
Errors 10850, 10860 and 10861 are wrongly reported as EX_USER, they should be EX_INFO.
When migrating from Adaptive Server 15.5 or earlier to 15.7 using the sybmigrate utility and the 'quoted identifier enhancement' option is enabled on the target 15.7 server, an error message similar to "The specified object 'sp_migrate_db_lockprom' is not found in database 'testdb'." may be reported.
When auditing is enabled for login events, then an incorrect audit message appears in the audit trail for Extended Encrypted Password login protocol and for Extended Plus Encrypted Password login protocol to describe the password encryption used.
When network password encryption is used with the UTF8 character set, auditing is enabled for login events, and the password specified is longer than the ASE limit of 30 characters, then the client connection is closed instead of reporting the generic login failure error and no audit event with the cause of the failure is placed in the audit trail.
DBISQL removes ASE-related preferences when closed without a connection to an ASE.
For data caches using strict replacement policy, the run_replacement and config_replacement column values are incorrect in the syscacheinfo and syscachepoolinfo views.
In rare circumstances the message "Infected with 11" together with modules like ex_printmsg(), senderror() followed by spid termination may occur when SET STATISTICS TIME is enabled with a create proc referencing an existent temporary table and a particular Sybase supplied traceflag is enabled.
ASE on Solaris platforms reports 694 or 823 errors along with messages like "kernel sddone: write error on virtual disk 3 block 145536", "kernel sddone: 16384 bytes passed, 0 returned on write for virt disk 3 block 145536" or "kernel sddone: read error on virtual disk 6 block 6812351", "kernel sddone: I/O error". There are no corresponding disk i/o error messages reported in the Operating System logs.
ASE may raise error 8201, when number of open object, indexes, partition is insufficient.
A 7153 error, "Premature end-of-message received while reading TEXT data from network. Host program may have stopped responding or may be sending TEXT data containing a partial character at the end.", may be reported in the error log during the replication of a table that has compressed TEXT/IMAGE data.
ASE running in Threaded kernel mode does not generate shared memory dump if CSMD condition is encountered by a non-engine thread.
For an LOB column which is defined both in-row and LOB compressed, if the column is marked as replicated, when replicating the LOB data from this LOB column, the data selected from the target table may be corrupt.
A 7975 error, "Page number < number > is out of range for this database master", will be reported when running DBCC TEXTALLOC on a table with TEXT column.
High rprocmgr or rpmctrl spinlock contention may be observed under heavy workload of dynamic SQL or stored procedures.
UNION query (using MERGE or HASH UNION operators) against semantically partitioned table may fail to remove duplicates when executed in parallel mode with constant literal search arguments for predicates on each side of the union operator.
When an INSERT query attempts to insert an explicit value for a TIMESTAMP column a warning will be reported, but query execution continues. If the explicit value for the TIMESTAMP column is the result of evaluating an expression, and the query plan for the INSERT is re-executed (i.e. because the statement cache and literal auto parameterization are enabled or because the INSERT query is part of a stored procedure), memory corruption may occur, leading to various stack traces.
Installer fails to optimize ASE if user chooses to configure an 16k page size ASE.
Replication Agent may incorrectly send "remove table" to Replication Server when a stored procedure is removed from the schema cache.
Permission is denied when a procedure is executed as a RPC event not as a LANGUAGE event.
The message "current process infected with 11" in the module 'bufpindirty' together with a stack trace which includes the modules 'xact__rollbackxact' and 's_execute' may be reported in the error log if a SELECT INTO statement involving a big number of tables (more than 96) is aborted with a 226 error, "SELECT INTO command not allowed within multi-statement transaction". If after that a ROLLBACK TRAN is executed by the same session a 6203 error "Unable to do cleanup for the killed process; received Msg ..." may also be reported in the error log.
In rare circumstances, the message "current process infected with 11" together with a stack trace which includes the module 'jvmh_utl_printn' may be reported in the error log when ASE attempts to bootstrap the COTS JVM. This will result in Java operations being impossible to execute.
After DBISQL receives an out of memory error when retrieving the rows of a large resultset, it fails to cancel the underlying query. Instead, it allows the query to run to completion and hold locks and other ASE resources for the duration.
After loading a dump from ASE version 15.5 or lower into a ASE version 15.7 server, when the server is first onlined, a signal 11 may be encountered followed by a stacktrace involving the modules 'ptnc_get_ptncond', 'ptnc_fill_ptnvals' and 'ind_ides_init'.
In ASE Cluster Edition, if a node crashes in the middle of DROP DATABASE, a signal 10 may be reported when the crashed node re-joins the cluster.
A 9830 error, "Compression version 0 is invalid for table 'Invalid pointer param number 3, pointer value 0x11' (ID < tblID > ), database < dbName > while compressing page < pageid > .", may be reported when doing a deferred INSERT into a table having page compression enabled.
The 'no_DBISQL' variable is missing in ASEIP.bat
The error message, "symbol _port_dispatch: referenced symbol not found", may be reported on Solaris 11 platforms. The '_port_dispatch' function call is no longer supported on Solaris 11 and is replaced by 'port_send'.
ALTER TABLE MOVE PARTITION can fail if the destination segment does not have enough space. However the transaction does not get rollback successfully. The SYSPARTITION update does not get rollback.
Creating a replication index, using sp_setreptable, sp_reptostandby or ALTER TABLE, on an in-row LOB column whose maximum in-row length is larger than the maximum allowable size for an index fails with error 1903.
ASE may encounter a SEGV after issuing DBCC STACKUSED and the task issuing DBCC STACKUSED may hang.
DBISQL login with expired password can fail if invalid entries are present in the sql.ini
In diagserver, the message "current process infected with signal 11" together with a stack trace which includes the modules "mda__read_lock_table" may be reported in errorlog when the monLock MDA table is queried.
Shared locks, which are blocked by demand lock on an earlier lock request, cannot be requeued after the blocking process is killed.
sp_helpthread shows a value of 0 for TargetSize in 15.7 ESD1 SMP.
Optcriteria 619756 (enabled in optlevels ase1503esd4 or ase_current) could generate a SEGV and/or assertion failure. The stacktrace associated with the SEGV will typically include references to optimizer routines. In a diagserver, the following assertion (eqc- > eqcIsUnrefed() || OptDuringGtuLocalGroupingPrimingHack()) may fail.
When a table has some compressed LOB columns and is marked as replicated, slow BCP'ing data into this table will hit 7105 error and stack trace.
In Threaded kernel mode, if an ASE task which is doing remote I/O is killed more than once, a SIGSEGV followed by a stack trace that includes the module np__conn_doconnect() may occur and the dataserver then crashes.
In rare circumstances, a signal 11 may be reported in the errorlog with a stack trace involving the sort module if a reformatting query plan is used.
When increasing the parameter 'number of disk tasks' on ASE on the IBM AIX platform, a stack trace and/or error 823 may be reported and the server becomes unresponsive.
ASE on Linux platform cannot start second or subsequent disk controller threads.
In rare circumstances, a 2601 error, "Attempt to insert duplicate key row in object 'syspartitions' with unique index 'ncsyspartitions'", may be reported when creating or partitioning a table, some time after the creation of a table within the database was recovered (either after an impolite shutdown, or by LOAD DATABASE or LOAD TRAN).
The error message, "Heap memory not freed", is printed in the error log during character encoding compression.
The message, "timeslice -501, current process infected at 0x < address > ", in the module 'cm_writelogdes_upto_sequence' together with a stack trace which includes the modules 'cmcc__wri_bcmt_proc' and 'bufwritedes_log' may be reported in the error log when ASE flushes a syslogs page on disk while downgrading the physical lock held on this syslogs page in a 2 node Cluster Edition setup.
After upgrading to ASE 15.7, DDLGEN may generate incomplete "lob_compression=" clauses in 'create table' DDLs.
The ASE server is unable to online an additional engine in Process kernel mode if Kerberos is enabled. The engine online will fail with the message, "engine < engine_id > , os pid < os_pid > exited" reported in the error log.
If the server is configured using multibyte charset, READTEXT with 'using chars/characters' option from compressed TEXT/UNITEXT/IMAGE column may encounter a signal 11 followed by a stack trace that includes the module txt__position() and readtext().
Executing sp_monitorconfig 'all' will display 'kernel resource memory' twice but fails to display 'compression info pool size' information.
In ASE Cluster Edition, error 941 may be reported when executing sp_configure "current audit table".
In situations whereby there may be existing in-memory identity information, it is possible for a insertion into a table with an identity column to produce a duplicate identity value after dump\load. This may be specific to a change in server-level 'identity grab size'.
A 712 message, "There is insufficient heap memory to allocate < number_of_bytes > bytes. Please increase configuration parameter 'heap memory per user' or try again when there is less activity on the system.", may be reported when creating an index on a table with a large number of partitions.
When an insensitive cursor is declared over a result set that contains at least two LOB columns and one of these columns is an in-row LOB with at least one result value actually stored in the row, a SIG BUS may be raised when the result set is FETCH. This error only occurs on platforms that require memory address alignment.
In ASE Cluster Edition, if create database reuses the dbid of a dropped database, it may hang due to blocking on some locks.
In rare circumstances, the message "current process infected with signal 11" together with a stack trace may be reported when involving compressed tables.
In rare circumstances, ASE may shut down with the message, "Stack guardword corrupted" reported in the error log when running queries with LIKE clauses on LOB columns.
Preupgrade may report an error that huge master device size is needed for upgrade to 15.7 ESD#2.
When a primary key contains IDENTITY columns, INSERT INTO and SELECT statements may create bad LTL causing the RepServer to report the error message, "Incorrect syntax near the keyword FROM".
Update statistics with sampling can fail to correctly write the sampling percent to sysstatistics, optdiag\sp_showoptstats will report zero. This is a diagnostic issue only, sampling is performed during the update statistics command.
The message, "current process infected with signal 11", together with a stack trace involving the functions keclearown(), uppopaffinity() and threadpool__update() may be reported when number of threads in a thread pool is changed, for example from 2 to 1 via "alter thread pool < thread pool > with thread count = 1" under ASE threaded mode.
The error, "Page Verification Error: Page mismatch. Page # contains # in page header", may be reported in the Backup Server log file if the database has two fragments on a database device where first logical page of the first fragment is smaller than the first logical page of the second but the first virtual page number of the second fragment is smaller than the first virtual page of the second.
ASE logs the message "Expected object passed to utGetOsId is not of type thread" and "Expected object passed to utGetThid is not of type thread" repeatedly.
Dumping on Windows using 'WITH COMPRESSION' clause can perform worse that the same command using the API library "compress::".
ASE reports message "kshpnetctlrWaitForCompletion : received error 8 for psn ..." when a connection gets killed abruptly.
The message, "current process infected with signal 11", together with a stack trace involving the functions hk__compress_page() and pg_modify_spacebits() may be reported durng HOUSEKEEPER garbage collection.
In installations with large number of user connections, the message "current process infected with signal 11" together with a stack trace involving the function sybcsi_default_entropy() may be reported from the housekeeper chores process or the job scheduler process.
A 3146 error, "ASE cannot load this database dump because the version in the dump is newer than the version in the database...", may be reported when a database dump from a Cluster Edition server is loaded in a newer SMP server.
A 706 message, "Process < pid > tried to remove PROC_HDR < prochdr address > that it does not hold in Pss.", may be reported in the errorlog when either the 'streamlined dynamic SQL' or 'enable functionality group' configuration options is enabled and recompilation occurs while a dynamic SQL query is being executed and an exception occurs during the recompilation.
In rare circumstances, the message, "current process infected with signal 11", together with a stack trace involving the functions collocate(), run() and s_execute() may be reported when executing a simple INSERT query with compatibility_mode enabled.
When multiple encrypted columns with the decrypt_default attribute are selected through a view, ASE may incorrectly return decrypted data on some of the columns.
Queries with OUTER JOIN on encrypted columns may not return expected rows from the outer table if the ON clause is on columns with decrypt_defaults and there is a where clause.
In rare circumstances, the message "current process infected with signal 11" together with a stack trace that include the modules collocate(),LeSubst() and getnext() may be reported when compressed tables are involved in the query.
A signal 11 may happen in CgpPop::CgpInplaceUpdateVars() for an INSERT...SELECT...ORDERBY query when the insertion involves RI constraint checking and one of the ORDER BY columns has a different datatype from its destination column (so that an implicit convert will be added) or the ORDER BY column is an expression.
In rare circumstances, the message, "Could not locate the before image for the INSERT log record", may be reported in the error log when replication is on the page-compressed table.
A 5851 error, "Configuration Error: Configuration file < config_filename > has an unknown format on line < line_num > .", may be reported during ASE reboot.
sp_downgrade prepare with 1502/15.0.2 as the @toversion reports "'18.104.22.168' is not a valid downgrade version" message.
Remote Backup Server name is now allowed to be specified as 'hostname:portnumber'.
On AIX plattform, " < num > suspect conditions" may be reported by DBCC CHECKSTORAGE and Fault 100018, 100019 and 100015 may be reported by sp_dbcc_faultreport.
The message, "current process infected with signal 11", together with a stack trace involving the function norm3_get_node_sqlt() may be reported during DISTINCT operator code generation when a DISTINCT view/derived table is participating in an expression EQUIJOIN.
On a symbolic binary, the datachange() built-in may hit a signal 11 during the cleanup of the execution after hitting 1105 error due to space crunch in tempdb. On an FBO binary, it may execute unnecessary code during the cleanup.
The srvbuild/srvbuildres (UNIX) and syconfig.exe/sybatch.exe (Windows) utilities now require user to specify the System Administrator password when configuring a new ASE.
When the master database is setup for replication, the command 'ALTER ROLE... LOCK' succeeds but is not replicated and an error 11314 is reported.
There is no entry for entering backend server SA user password for creating a new cluster.
The message, "current process infected with signal 11", together with a stack trace involving the functions stat_mergestats_all(), sort() and cri_ptn_index() may be reported when a parallel CREATE INDEX specifies that 0 histogram steps should be used, effectively disabling histogram creation, and multi-attribute density creation.
In some situations, a signal 11 may occur in lock__get_heldlocknum if a deadlock or lock timeout event happens.
While running ASE, the time a task waits on a lock is incorrectly calculated by adding a overlapped time.
The ASE default task stack space is too low for default execution coverage.
The message, "current process infected with signal 11", together with a stack trace involving the function prot_view_decdef_cols() may be reported with a complex query when access is made through a VIEW and the accessed columns are encrypted with decrypt_default.
When using tables with 64 or more partitions and parallelism is enabled with max parallel degree more than 64, ASE will stack trace in bitstr::test() method.
When a DUMP TRAN is done on a non-existent database, the expected 911 error, "Attempt to locate entry in sysdatabases for database ' < database name > ' by name failed - no entry found under that name. Make sure that name is entered properly.", is not reported.
The message, "The Begin Timestamp= < ts1 > of the completing transaction in database < dbid > is lower than the Database Commit Timestamp= < ts2 > ", may be printed in the error log together with a stack trace that contains the modules dbts_commits_remxdes() and xact_end_session() when chained transactions are used, and ASE is started with the diagserver binary.
In rare circumstances, when replicating compressed LOB data, the replicate ASE may raise error 7153, and the Replication Server may abort the connection.
When running UPDATE INDEX STATISTICS or UPDATE STATISTICS for a specific column, the leaf level of an index is scanned to get the values of any non-leading columns (if available). It is possible for ASE not to select the most efficient index to do this operation if the column exists as a non-leading column in more than one index.
When using ASE Multiple Path Replication with distribution per connection and while replicating LOB data, the Replication Server DSI may shut down itself with either error 5139, "DSI encountered an error when processing an rs_writetext command for object ' < Object Name > ', text column ' < Column Name > '. This rs_writetext command is not marked as the first chunk for the column. However, there is no previous rs_writetext command that is marked as the first text chunk." or error 5147, "DSI encountered an error when trying to do writetext operation for object ' < Object Name > ', column ' < Column Name > '. The data row that the writetext command is associated with is missing.". Under rare circumstances the Replication Server may also shut down itself.
sybdiag shows "14000 - Interfaces Load Error" if non-default interface file is specified with -I option in the command line.
A 7153 error, "Premature end-of-message received while reading text data from network. Host program may have stopped responding or may be sending text data containing a partial character at the end.", may be reported when replicating in-row LOB data that is marked as "replicate_if_changed" and the UPDATE statement does not modify the in-row LOB data.
ALTER TABLE < table name > split partition < partition name > into < partition values > or ALTER TABLE < table name > move partition < partition name > to < seg name > will encounter 691 error when db option 'full logging fopr alter table' is enabled.
When upgrading from a pre-15.7 ASE version that does not support compression, RepAgent ASE 15.7 incorrectly assumes that compression is on for uncompressed tables.
The message, "Current process (0x < value > ) infected with signal 11 (SIGSEGV)", in the module 'kbfalloc' together with a stack trace which includes the modules 'ra__process_xtext' and 'ra__process_txtblk' may be reported in the error log when replicating large, compressed TEXT columns.
In rare circumstances, when replicating compressed IMAGE/UNITEXT LOB data, the replicate ASE may raise a 7153 error, "Premature end-of-message received while reading text data from network. Host program may have stopped responding or may be sending text data containing a partial character at the end", and the Replication Server may abort the connection.
Following an abnormal termination of ASE, the recovery analysis and redo passes can be unnecessarily long in recovering large databases that had a very large number of different pages flushed to disk since the last checkpoint prior to the termination.
When an installation is created by ASE version 15.0.3 or before, then upgraded to 15.5, and then further upgraded to 15.7 or later, then if the upgraded server is downgraded back to 15.5, the booting of the 15.5 server may encounter a segmentation fault. This fault prevents recovery from completing.
The message, "timeslice < value > , current process infected at < value > ", in the module 'io_submit' together with a stack trace which includes the module 'bufawrite' may be reported in the error log when ASE is started in threaded mode and while it is creating a large database.
Sometimes a 619 error, "A deferred update was requested but the query is not one of INSERT, DELETE or UPDATE", may be reported while executing WRITETEXT command on an in-row LOB column.
A 301402 error, "Backup Server: 22.214.171.124: Logical page < page_number > is not in the given database map.", may be reported by the Backup Server during the execution of LOAD WITH VERIFYONLY on a database whose size is bigger than 2147483647 pages in an ASE installation that uses page size greater than 2K page.
ASE-UA Role Mapping (Roles Folder), UA Security Modules (Server Properties Agent page) and UA Attributes/Operations (Server details panels) are no longer supported due to upgrade to SCC 3.2.7.
When an in-row LOB column which has both data compression and LOB compression properties defined has its schema changed by setting the LOB compression property to not compressed, ASE following a SELECT query execution may return incorrect results for this column for data that was inserted into this column before the column schema change occured.
Syntax is added to the UPDATE STATISTICS command to allow printing of progress messages while the command is running. The new syntax is UPDATE STATISTICS < tablename > ... with "print_progress"= < 0|1 > where 0 = disable and 1 = enable. The default value is 0.
Create cluster using sybcluster failed with message, "Unable to retrieve the environment variables to run the Sybase tools".
A 1525 error, "Sort Failure. Rollforward of sort encountered sort descriptor timestamps out of sequence. Old timestamp in log: < old_ts > . New timestamp in log: < new_ts > . Timestamp in sort descriptor: < sort_ts > .", may be reported when loading a transaction dump that contains a CREATE INDEX command that was executed using parallel sort. ASE should have disallowed a transaction dump to be taken if it contains a parallel index creation.
In rare circumstances the server can be shutdown after an error in the module dbts_commits_remxdes() followed by a message 'The Begin Timestamp= < n > of the completing transaction in database # is lower than the Database Commit Timestamp= < n > '.
When Replication Agent is enabled to use the multipath distribution per connection model, and if the option "ddl path for unbound objects" is set to "default", it is possible that DDL commands are replicated to the incorrect path.
Data may be corrupted during page compression if there is a compressed in-row LOB column in the table.
A 1163 error, "The space allocated to the sysgams table in database 'tempdb' (dbid = 2) must be at least 16 pages.", together with a stack trace which includes the modules 'pggam__srchonegampage', 'pg_allocate_syslogs' and 'crdb__logtrunc' may be reported at ASE startup if the system tempdb is bigger than 63GB and has dedicated log devices which are placed out of the 63GB range.
While running a batch of ADD or DROP constraints out of which a few fail due to valid reasons, the subsequent ADD/DROP constraints may fail with the error 4956.
The error messge, "Page Verification Error: Virtual page < page_num > in device < dev_num > is out of database range." may be reported in the Backup Server error log as a result of a LOAD DATABASE command.
Network Controller loops infinitely in trying to lock a spinlock.
In rare circumstances, when the ASE configuration parameter "disable varbinary truncation" is set to 1, a delete from a DOL table which has a unique index which has varbinary columns may hit a 644 error, "Index row entry for data row id (Pageid, Rowid) is missing from index page < page id > of index id < index id > of table < table name > in database < db name > . Xactid is ( < Transaction Id > ). Drop and re-create the index."
Using sybcluster to add an instance in interactive mode failed.
ASE is not able to boot large number of engines in Process kernel mode as Spinlock allocation fails.
Permissions through roles granted to non-default login profiles can be lost or reduced if there exists a default login profile that has not been granted all the roles that the non-default login profile has and a permission changing event (such as a grant or revoke) occurs in the server.
The ASE errorlog may contain many instances of the kernel message "Expected object passed to blkioDescriptor is not of type BLKIO,..."
In rare circumstance, when replicating an off-row compressed LOB column, the replicate ASE may raise the 7153 error, "Premature end-of-message received while reading text data from network. Host program may have stopped responding or may be sending text data containing a partial character at the end." and Replication Server will abort the connection.
If there are more than 32 columns used in an UPDATE STATISTICS command, either explicitly or implicitly through the auto_temptable_stats feature, then there is a possibility of a stacktrace.
In some cases, DBCC CHECKTABLE may report error message 7948 after a clustered index is created on a compressed APL table which contains only 1 row.
Complex queries (often with a large number of group-by columns and a large number of joins where generally each grouping column belongs to a different "joined" table) may SEGV after generating a 331 error terminating the session. The 331 error is raised because the query has exhausted the sort id namespace.
In rare circumstance, error 1204, "ASE has run out of LOCKS. Re-run your command when there are fewer active users, or contact a user with System Administrator (SA) role to reconfigure ASE with more LOCKS.", and error 5619, "Adaptive Server internal error. Could not obtain count of in-doubt transactions for space reservation.", may be reoported when running BCP IN/INSERT/UPDATE on table with TEXT/IMAGE/UNITEXT columns marked replicated.
Targeted CR List for ASE 15.7 ESD#2
The purpose of this document is to help Sybase customers obtain a general idea of potential bug fixes in future Adaptive Server Enterprise (ASE) EBF releases. Please be sure to read the Disclaimer and General Notes below.
Disclaimer: This document lists the targeted (not committed) fixed CR list for the release listed above. Please be aware that the purpose of this posting is solely to provide you, our customers with estimated release dates and targeted CR lists. Sybase does not commit to releasing EBFs on the specified dates or to including the CR fixes in the said EBF. While every effort will be made to meet the said targets, changes can occur at any time. It is also possible (although unlikely) that Sybase may decide not to release an EBF that was previously scheduled for release.