Sybase数据库技术，数据库恢复专家

This issue can specially be seen when users are migrating from ASE 12.5.4 to ASE 15.7.x

will be 1 less than the actual number of rows in these two system tables.

incorrectly which could result in wrong results.

option on compressed in-row LOB columns, dbcc checktable() may report various

kinds of index errors.

It enables Abstract Plan sharing between different users. Tables in the query with shared

abstract plan must be explicitly prefixed with the owner name or belong to the DBO user.

installed by the user or were previously migrated. Cannot proceed with migration of 'Users'".

It is more likely to happen when the two phase aggregate plan is also a parallel plan,

or when the query is also a GROUPBY query with a HAVING clause equal-comparing two of the count aggregates.

in database '<dbid>'. Please contact Sybase Technical Support." may be reported when

a 12.5.4 ASE database dump that was made with a 15.7 ESD#4.2 Backup Server is loaded

back in a 12.5.4 ASE.

so customer can be sure that they have enough space when they are ready to install.

cursor) involving a list of over 128 OR conditions it is possible that the delete will fail

with error 592.

This may also occur when issuing the sql language comnmand 'delete where current of'

against an updatable cursor.

result may be returned in rare circumstances involving a derived table or view containing

a union all with the aggregate operating on a constant in the select list on at least one

side of the union all.

and followed by a stack trace that includes the modules bt__handle_LAdone(), bt__qualpage(), bt__scan_massoffset() and bufoffset().

lio/cpu/locks/elapsed_time/tempdb_sapce similar to resource govern.

sp_configure 'quoted identifier enhancements', 0

execution of Ct-library calls which may result in undesirable behavior.

of a derived table (or view) column and a string value. Also involving a correlated sub-

query within the derived table definition it is possible for the optimizer to generate a

plan which may lead to an unexpected error 249 (or similar) dependant on data.

of a reserved double or triple keyword phrase it is possible for an error 156

error to be reported along with error 11934 and for the REORG REBUILD command to abort.

it is part of ASE_CORE license

not just a warning message

in one batch.

CHECKDB or DBCC CHECKTABLE is run on a database/table which contains variable length columns previously.

Exiting", is reported when a cumulative dump is used as a metadata file for sybdumptran.

which are created on a shared disk cluster ASE server, INSERT task may not respond even if it is

cancelled (by CTRL C) when run with concurrent INSERT tasks. Meanwhile, lots of

1205 error messages, "Your server command (family id <value>, process id <value>) encountered

a deadlock situation. Please re-run your command", will also be seen in the error log.

identity column will be populated with correct value but the global variable @@identity will

remain unchanged.

(pseudo code)

UPDATE row

if (no rows found)

INSERT row

if (duplicate key)

UPDATE row

if (no rows found)

ERROR out (assumption this should not be possible)

The new feature is enabled when the new configure option "donot skip uncommitted insert" is set as 1.

'oldest_active_transaction_pct', 'stp_span_pct' for databases having mixed log and data segments.

in a multi-statement transaction may sometimes not qualify all rows, and the DELETE / UPDATE

may leave some qualifying rows unaffected. This is most likely to occur when the affected table is

bound to a very small cache, usually of the minimum size.

To reproduce:

Use sample OOM to generate an XML model

or

1) create an OOM

2) add a Class with several attributes

3) enter some text in the Description of the Class

4) Generate XML model; on the Detail tab select 'Generate Attributes as 'Attributes'

In the resulting XSM, the attributes have not been generated for the Class

"The UPDATE/DELETE WHERE CURRENT OF failed for cursor '%.*s' because the curso

r is read only."

The CIS query looks like:

"UPDATE PROXY_TABLE ... FROM LOCAL_TABLE, PROXY_TABLE where ..."

For the invalid query plan which caused 7732 error, it has join order proxy_table -> local_table,

ORDER BY join_column clause is added for the partial query sent to remote server,

remote server declared a cursor using a statement looks like:

SELECT ... FROM ... ORDER BY ... FOR UPDATE OF ...

This issue is related to a known CR 652032: "Allow positioned updates when ORDER BY is used".

Please force a valid plan via a valid abstract plan or force join order via 'set forceplan on' to work around this 7732 error.

non-unique parameter or with an individual parameter

on ASE 15.7 SP100 release

(1) The configuration option "streamlined dynamic SQL" is enabled, either directly or by enabling functionality group and

(2) A dynamic prepared INSERT or UPDATE statement is executed on a table that has been created with deferred_allocation and

(3) The prepared statement has been executed successfully at least once and

(4) The target table that is inserted/updated is dropped and re-created as a defered_allocation table and

(5) The dynamic INSERT/UPDATE query is re-executed from the same ASE connection.

be corrupted resulting in a self-referencing ptr in a names struct.

This can result in a process allocating the entire procedure cache

under cpynameslist(). The process cannot be killed requiring a server

reboot. A check for a self-referencing names struct has been added in

cpynameslist()and the overloaded error 7733, state 2 added if such a ptr

is found. This error can be used for configurable memdumps. This check

for a cyclic ptr reference is enabled under trace command line flag 7728.

the update to fail reporting error 614 and for a stack trace to appear in the ASE error

log along with a message 'Invalid DOL data row pointer' and Error: 614, Severity: 21,

State: 12

'Adaptive Server accessed a row that has an illegal length'

1. If the login is created ignoring login profile, the "create login..." command lacks the clause of 'login profile ignore' or "sp_addlogin" is used.

2. If the login is created with login profile and the login profile has blank default database.

logical page '...' in sdes '0x...' kept buffer pool for object '...'.

a database which cannot be currently accessed.

during key movement.

be upgraded to hard fault after running DBCC CHECKVERIFY.

during key movement.

The low configuration/performance gains due to the use of transactional memory may be interesting to the customer though.

passes the return value of the second call to the first call.

for a long time. A process might be sleeping and waiting for an EVENT to happen, if the related code doesn't check attention, then it's possible to sleep forever if the EVENT doesn't happen due to some bug or some other reason. Reboot to clear the hung process is not easy for some production systems since DBAs may need to schedule reboot time.

by Certicom in FIPS mode are supported by OpenSSL.

DBCC CHECKTABLE.

in master..monRepLogActivity are only maintained when sp_sysmon is running in parallel at the

same time.

"Message: 10331, State 2, Severity 14 -- 'Permission denied, database master, owner dbo. You need the following permission(s) to run this command: MANAGE ANY LOGIN. "

The error is mapped to stop replication.

Msg 208, Level 16, State 1:

Server '<servername>', Line 26:

<#temporary_table> not found.

This happens if the same trigger is defined for both INSERT and UPDATE:

"create trigger T on TAB for insert,update as ..."

'not null' and 'not materialized' in this table, update (index) statistics on this table may

hit stack trace and SIGSEGV at decompress__column_by_colinfo().

headers (...) does not match memory usage count in Pss (...) for server

process id ..." may be reported in the error log when the configuration

options "enable functionality group" and "plan sharing" are active and

the execution of a prepared statement is interrupted due to an error

that caused the current transaction to be aborted.

allocation, server might hit error 8419 as :

Error: 8419, Severity: 20, State: 3

Could not find index descriptor for objid 240003886, indid 0 in dbid 19.

issue in the middle of update syspartitions table.

if it's run after shrink database operation.

1. If the length of hostname and port number together are equal to or more than 30 characters, login redirection fails.

2. SAP ASE runs into different issues when server options are changed using sp_serveroption for HADR_GROUP class(17) entry in sysservers.

"Execution of rule check_jsc_state failed because of errors parsing the source text in syscomments during upgrade. Please drop and recreate dbo.check_jsc_state"

overflows.

This may lead to error: "12316: Attempt to garbage collect a deallocated data page 13407 in table 'table1_PDR', database 'onlinedb'. Aborting the transaction" could be seen.

Earlier used to print a message for every cache-partition in the cache, now it will be printed only for 1 cache-partition. This will avoid flooding messages in the errorlog.

'<index name>'", may be reported during the load of an 12.5 database in a 15.7 dataserver.

statistics where column histograms are out-of-range. An UPDATE STATISTICS option is

provided to implicitly generate new histograms which include the out-of-range values

for a user-specified column.

<value> to a <type> field ." can be seen while running stored procedure sp_spaceusage on large table.

Note that the ase_current optimization level will not enable this functionality. In order to enable this functionality, explicitly execute "set cr750883 on", or define a user defined optgoal that includes this optcriteria, or enable trace flag 12801 on server startup.

"Warning: '<large device physname>', is a file system device. A system failure may cause data loss if the dsync option is set to false."

Pages deallocated during merge partition were getting allocated to same or other tasks even before

merge partition is over.

The optimization is enabled by optcriteria.

If traceflag 2792 is ON HK_GC will wait there, which may then cause an undetected deadlock.

with "alter thread pool', and then created again, Security Control Layer initialization fails due

to exhaustion of memory resources with errorlog message:

00:0061:00000:00000:2013/11/07 00:41:47.50 server Security Control Layer's memory allocation routine returned an error.

dropped during query optimization" may be reported when SET ANSINULL

statement is used inside a batch containing other statements.

The label for the third radio button is changed from "Cached IO (Data sync off)" to "Buffered IO"

E. 2013/12/02 08:55:31. ERROR #1028 DSI EXEC(106(1) db_srv2.ads_main) - dsiqmint.c(4358)

Message from server: Message: 249, State 1, Severity 16 -- 'Syntax error during implicit conversion of VARCHAR value 'Dez 1 2013 11:00:23:020PM' to a DATETIME field.'.

(here, 揇ez� word (in datetime value 揇ez 1 2013 11:00:23:020PM�) is the short name for December month in German)

This may also impact other "date" data type.

PCC SP60 is released with OCS SDK 15.7 SP122.

PCC SP60 is released with jConn

jConnect (TM) for JDBC(TM)/7.07 SP120 N-OFF (Build 26972)/P/EBF21999/JDK 1.6.0/jdbcmain/OPT/Thu Nov 7 06:04:54 PST 2013

PCC SP60 is released with ODBC

c:\DataAccess\bin>odbcversion -version

15.7.0.401

No Power Designer is shipped with this image, as there are no updates to PD. Please continue to use your existing Power Designer installation.

1. The SFU query contains a TOP clause or a 'set rowcount' command was previously executed to limit the number of rows returned by the SFU query and

2. The number of rows that qualify for the SFU query exceeds the limit imposed by the TOP clause or 'set rowcount' command and

3. The optimizer chooses a parallel query plan to execute the SFU query.

If the stack overflow backout code uses too much stack, the backout process can cause stack guard word corruption resulting in server shutdown.

This CR removes the limitation on number of frames in the stacktrace and if the stack overflow backout code use of the stack reaches within 4K bytes of the stack boundary, it then initiates forcible termination of the task to prevent stack guardword corruption.

The default value of 'stack guard size' is increased by 4K bytes to retain earlier level of stack availability.

Traceflag 3687 provides a stacktrace when a transaction is aborted due to insufficient stack space that can provide additional diagnostics on stack usage by the query. Traceflag 3686 disables this change. This change is not applicable to the Microsoft Windows platform.

object 'object name' (object id), index 'index name' (index id), partition 'partition name' (partition id).

This is an internal system error. Please contact SAP Technical Support." may be reported with stacktrace when

inserting data into TEXT/IMAGE/UNITEXT columns.

when a list partitioned table is the inner table of a parallel outer equi-join and the

join is on the partition column.

The same stacktrace could also happen for a range partitioned (instead of list

partitioned) table when the upper bound of the partition condition is not MAX.

truncation point set, it indicates to turn the secondary truncation point off using older server, and

when rebooting with the older server an error 950(Database 'master' is currently offline. Please wait

and try your command again later) will be raised.

to fetch logical page <page id> from cache <cache>. Page belongs to database <db>, object <object>,

index <index>, partition <partition> and not to database <db>, object <object>, index <index>,

partition <partition>.", may be reported when running 'SELECT...INTO...FROM ... WHERE condition1

or condition2 or ...' command on it with concurrent DELETE/UPDATE/TRUNCATE TABLE commands.

There are still external logins for the server 'mtrs$999'.

references a view or derived table that contains a union-all that terminates

a union statement may generate a 264 error (duplicate column id) if the

identity value is implicitly generated (i.e. - not explicitly provided in the insert's value list).

update/delete statement.

or SET STATISTICS PLAN_HTML ON the runtime execution statistics

reported for some queries could be inaccurate when the plan displays

some operators like "HASH JOIN PROBE", "HASH JOIN BUILD", "SORT (GETSORTED)"

or "SORT (SORTBUILD)".

Could not find partition descriptor for objid 30, indid -1, ptnid 30 in dbid <dbid>.

Column sysobjects.sysstat4 has length '0', should be '1'. The sysobjects row is for database '<db name>' (db id), object '<object name>' (object id).

shutdown if some other process is listening on the xpserver port. ASE shutdown

may also hang if some other process is listening on the xpserver port.

.

on the partition column(s)

If an invalid version record is found, initialize the history file after copying the file to a back up.

Previously, when creating a global index on partition table, if there were any empty partitions, then the index would be created in serial.

target table. Dbcc checkalloc or tablealloc may report Msg 2546.

Previously, a user defined optgoal needed to be defined to enable these optcriteria (or trace flags 16989 and 16991 respectively).

After hitting 12212 error SAP ASE terminates the connection and logs a stacktrace in the SAP error log.

Error 12212

" Internal Error: The partition ID (<ID>) of the lock does not match the partition ID (<ID>) of the new request on this lock. dbid=<DB_ID>, objid=<OBJ_ID>, pageno=<Page number>, rowno=<Row Number> locktype=<lock type>".

a cached language cursor.

Frequency: Occasionally

Server Enviornment: On heavily-loaded

Platform: All

Symptoms: Server hangs when using UPDATE INDEX operator in showplan and an Error 1105 occurs

subsequent commands to get fail with error 8601 "Process <n> tried to keep the PSS for process <m> when it already had the PSS for process <p> kept."

predicate is on two columns with NULL values and the final plan is using a special

OR strategy plan for the OR predicate. An example of such an OR predicate is "t1.col1 IN (1,2,t2.col2)".

-Kerberos

-LDAP user authentication

-PAM user authentication

Also, in scrollable table mode, previously Messages panel limited the number of messages to 1000 items. This limit has been removed.

/* 12319, EX_CMDFATAL, 1

** Fatal internal error encountered during deferred dml on table '%.*s' in database '%.*s'. All rows that qualified during the first phase were not visited during second phase of dml.

*/

the INSERT into temp table with select distinct UNION statements

"Sort (Build)" and "Sort (GetSorted)" for the splitSort

operator. These strings generate an XML parser error

when running a query that qualifies XML using xmltest.

"server Opening Master Database ..." and

"kernel The secondary truncation point is turned on in the master database. Please use your older ASE version to turn the secondary truncation point off before using this ASE version".

useful for tables and indexes with large number of partitions.

A client application may also raise an error message "ct_results(): protocol specific layer: internal Client Library error: There is a tds state machine error. An illegal tds token sequence was received.".

information on the time spend in the transaction commit processing when using streaming replication

in (near) sync mode.

an m_join property fails). In the case where the feature to eliminate

substitution elements that appear only as simple sargs is enabled, this

results in a SEGV in LeVtuple::LeVtCopy under LeGroupSortedOp::_LeOpNext.

Tcf 16938 can be used to disable the substition scalar elimination.

pageid values are being represented as signed 32-bit integers even though

large databases may generate pageid values > 2,147,483,647

optimizer timeout due to exceeding procedure cache limit.

Provide an application\utility replacement for SAMreport.

Today, license server usage can be monitored manually by running "sysam status" command for the license server. However, this does not provide the same coverage as SAMreport.

For example, in large install bases, PowerDesigner customers typically use floating licenses (FL). Therefore, they reserve a number of entitlements for specific groups and\or users in order to ensure that Power Users have access to the software when needed.

Also, they need to monitor the package usage. Thus, they need the historical data.

ASE may send badly formatted data. The bad format is reported by the

client message: ct_fetch(): protocol specific layer: external error: There is a tds protocol error. An illegal token was received.

1) Predicated Privileges are not being automatically re-normalized after being loaded from a dump.

2) When a predicated privilege fails to re-compile, the exception handling incorrectly cleans up leading to 706 errors.

the module 'LeAlfaStatsCollector::LeWriteThStatsToCollector' may be reported

in the error log when a query is executed inside a stored procedure (or

the statement was cached) and the plan is executed in parallel while

some monitoring options are altered.

The rule condition includes the plldegree setting for the table's useful indexes (if any).

"plldegree" can be set for a table using the following command:

sp_chgattribute t1, "plldegree", 4

The rule is only active when the cmdline traceflag 16993 is on.

Also under the same traceflag, optimizer will not use a table's plldegree setting as

its useful index's plldegree when the index doesn't have the attribute set.

Please use system procedure sp_dbrecovery_order to change database recovery order, e.g.

sp_dbrecovery_order saptempdb, 1, force,strict

It will let ASE rebuild user defined tempdb saptempdb first after recovery system databases, then recovery other user databases.

This cr would correct the behaviour to make it CDS/BW compliant.

routines:EVP_DecryptFinal_ex:bad decrypt.'"

Customer asked to raise an error message instead of doing a hard termination with infected 11

This will not affect UPDATE STATISTICS or the serial CREATE INDEX command.

"Error 712: There is insufficient heap memory to allocate %ld bytes. Please increase configuration parameter 'heap memory per user' or try again when there is less activity on the system"

ALTER TABLE ADD NEW_COLUMN_NAME VARCHAR(column_length) DEFAULT ' '

may fail to insert the default value to the new column if database option 'allow null by default' is on. ASE may output 'Invalid column length' error to errorlog when it read a row having invalid column data.

For Business Suite on ASE installation. Process of customizing SAP application systems may alter SAP application system table which is stored in ASE database, e.g. 'SAPSR3.MARA'. If a SAP table stored in ASE has just been changed with new columns added during customizing procedure, and there exists invalid column length message and stack trace in the errorlog while accessing to this table, then please consider extract the data and reload it, and also to perform this during downtime.

For standalone ASE, please drop and recreate related table if it exists invalid column data to resolve the problem.

If a database has database option 'allow nulls by default' on, then may need to specify 'NOT NULL' to add VARCHAR column with space as default value via ALTER TABLE, e.g.

alter table table_name add new_varchar_column(column_length) default ' ' not null

it is executed using "INS_BY_BULK" optimization criteria, the target

table has several indexes and the system does not have any available

worker process.

that may appear in the error log, or other diagnostic outputs. The encoded code location can be

decoded into its constituent file name / line number parts using this asecodeloc utility program.

4-byte int value of the file name / line number where the error message is printed. Using a new

'asecodeloc' binary, this encoded value from the error log can be decoded to accurately locate the

source of the error resulting in the message to the log file.

in the module 's_setreturns' or 'LeEmitNoSndOp::_LeOpNext' together

with a stack trace maybe be reported in the error log when Job Scheduler

is started using the stored procedure SP_SJOBCONTROL and a 1204 error,

"ASE has run out of LOCKS. Re-run your command ..." is triggered.

for star join fact table hinted strategies if possible. Of course, this is

only applicable when parallelism is enabled.

the child thread will continue to drain it's plan fragment even when the pipe it is

writing to is closed leading to bad performance.

can raise Msg 712 insufficient heap memory.