March 9, 1996 Sybase Customers, The following represents problem reports still open and unresolved against System 11. This report is provided to aid you in identifying known problems, should you encounter any during the operation of this product. The problems reported are ordered by importance (high to low), and then by problem identification. Our development process specifies that System 10 bug fixes be included in System 11 products. There are certain circumstances where this is not possible, specifically when: a) Problem fixes made in System 10 missed the cut-off date for inclusion. That would be problems resolved past the System 10.0.2.4 Rollup date of November 11, 1995; specifically those resolved after January 19, 1996, b) System 11 code implementations made specific problem fixes made in System 10 unnecessary, c) Problem fixes made in System 10 were believed necessary, but presented high risk of destabilization to System 11. Therefore, they weren't included in System 11. At present, there is only one problem that is associated with item c); 61628. Current plans call for these problems to be included in a subsequent System 11 maintenance release. Sybase System 11 Release Team System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 78923 showplan output "Message empty." message, when connecting toSQL server by isql -J. If w/o -J option, in other words, access SQL server in english mode, showplan does not output "Message empty" message. 81045 ctlib.loc is missing under locales/us_english 81046 sql server stack traces when trying to retrieve rows from a table (looping with empty buffer) through front-end application. using ct -lib 10.0.2 81242 select datadd (day, isnull(NULL,0), getdate()) -- this command will crash the server. 81617 client errors generated on sql server w/7 or more engines and 300 users. Not seen with less than 7 engines or smaller number of users. 82463 when 11.0 server is booted with 4.9 device, only sa user can log in.Also, the server stacktraces when executing sp_who or a simple select from a user table. This prevents draining logs in replicated databases for upgrade. 82580 In altsort/alts003: Error 624 "Attempt to retrieve row from page via RID failed because the requested RID has a higher number than RID on the page. Followed by "timeslice -501 current process infected" and STACK TRACE: 82956 We see lots of deadlocks while running the TPC-C benchmark on NT platform. The GA version of the server behaves very well where the server generated on 2nd Jan comes out with lots of this deadlocks. Tested a server by backing out bugfix 73309 checked in 82977 In hsm_build_map() the device numbers (devidx) computed during the process of assigning the "shared" devices (rem_devs) is off by 1. 83144 zeta18_11 test hangs indefinitely until killed by KILLTIME parameter which is 8 hrs. Seen on HP, Solaris & IBM but not on OSF(where it fails with bug 82576). Occasionally, the test does not hang but aborts with bug 79833. Never completes successfully. 83147 alpha test hangs within an hour (less than 400 iterations) 83391 @@identity global variable does not report last identity value use properly. (see ~ts/cases/10058000s/10058825/dseinfo/Repro) This behaves differnently between system10 and system 11. 83401 select at isolation level 1 in busy_reader1 returns data in different order vs. 11.0.1 (compared with 11.0). /svrsundly_nttst1/tst_sql/960117_dly2/isolevel/busy_all/DIFF/busy_re ader.out Sybase Confidential Page: 1 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 83447 Running sp_configure on a vanilla server ( with default configuration) causes Message "701,17,3" not enough procedure cache 83456 s_code_generate() tests for SEQ_OUTPARAM and assumes a SEQ node even though a subquery node could be present instead, this could result in unpredicatable subquery results since R2T_OUTPARAM is incorrectly set in a subquery mainestep 83507 tmp_alldrop() calls drop_with_retry() with DROP_NOTFOUND_OK flag. drop_with_retry() returns TRUR (drop of temp table is a success), if this flag is set regardless of the error. This seems to be wrong. If this is correct, please add comments as to why it 84132 In the list of dboptions provided by "sp_option", the last two options 'trunc log on chkpt' and 'trunc. log on chkpt.' seem to be the same as there is no info in the docs to think otherwise. This is confusing and only one of them should be printed... 84375 procedure sp_columns, sp_databases, sp_sproc_columns, sp_stored_procedures, sp_tables has a shorter length in the column "remarks" when comparing with the canon. Run mcsp/[colm,dbs,prcl,stpr,tabl] to reproduce diffs. 84385 A language cursor is declared in a transaction isolation level of 3, with noholdlock, then it is attempted to open it at level 0. The error Msg 7792 return "...level 21 scan is being attemped..." while the canon have "...level 0 scan is being attempted.." 84390 Force plan does not effect the join order in stored procedure when using temp table. The temp table is always selected as first table in join order. 84430 The RVManager needs to check RVM_DBCC_PIA structure contains all entries. For a new dbcc cmd, simply adding a token in dbcc.h worked fine on solaris and hp800 but seg faulted on osf indicating that the code is picking up bad values past the structure. 84649 When doing alter table, SQL Server may( probably is very low, but ca n happen)incorrectly report that the column is already is already present when actually it is not.y 84758 960201 build:orderby/orby059:negative test using the same correlation more than once, "int_value" used 3 times. Test returns Msg 7348 4 times; CANON gives Msg 7348 3 times. 84807 960201: On NT & AXPOSF only. cgrbeta/76176/76176b: Cannot drop the database 'db76176' because it is currently in use. 84893 Test bufman/procs hang at "dbcc traceon(3604)" in buf580.sql. No error or unusually mesg in the log. The "sp_who" shown only an "insert" cmd running (This could be from bugid 79772). Same in rs6000. Sybase Confidential Page: 2 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 84894 4.1.2 "total size of all configured caches....is greater than the memory available for caches" in server log when boot with INVALID CACHE SIZE IN CONFIG FILE (negative cache size). 4.1.2 "Invalid pool size of ..." in server log when boot with INVALID POOL 84904 sp_cacheconfig in bufman/conf120 returns info on cache3, cache4 in different order than CANON, conf130 returns cache2 Wash Size as ****** vs CANON's x. 85016 At a point which user connections reach above 300 (application opens front end, each front end opens 3 user connections) 605 is encountered on temp db. stack trace available. 85069 System stored procedures are inconsistent and uninformative in the return code they return. The Peoplesoft app. needs to react differently depending on why sp_addlogin fails. Parsing the message text is not an option. 85272 ksource/ksndmsg.c declares functions for tli and socket under the ifdef of TCPIP_TLI and TCPIP_SOCKET. But the header file kinclude/ksndmsg.h file uses these functions under #if SVR4 ( for TLI) and #else for socket. 85274 endupdate() calls tally_row() to update the row counts in the oam when a transaction completes. This requires several expensive trips to the buffer manager to get the oam page, dirty it and unkeep it. There is no need to maintain the row count for ...w 85444 Server stack traces with 806 error when trying to load a 900 warehouse TPC-C database at Sun facility. The load in done with logging disable using a db-library load program. conn_hdlr- >sequencer->s_execute->bulk_main->bulk_normal->insert->cinsert- >getpaee 85473 [ASID:joins2/joins_0060/10] --- when the clause "where = is invoked in a query datatypes of the colums can not be converted to match each other. msg 257 returned saying 'CHAR' to TINYINT is not allowed 85476 [ASID:sp_configure11b/configur_1300/15] -- server returning "0" instead of error message 17260, see also bug: 80267 85479 [ASID:sp_configure4/configur_0220/1] In cases 1,2 instead of 17413 msg server is returning "0" while sp_configure devices option is invoked.case3, 17414 msg not seen while configureing default language, but server returning "0" 85686 The section for "Empty Page Flushes" in sysmon_index is repeating information reported on the previous line. No new data is retrieved from #tempmonitors. A select statement must have been lost somewhere during integration into the codeline. Sybase Confidential Page: 3 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 85696 sp_figure reture extra lines in Group Languages: 'disable character set conversi' and Group Network Communication: 'allow sendmsg', 'sub_sendmsg port number' 85780 [ASID:auditing5/auditing_0240] --- local server is being treated as remote server and every remote procedure call to this server is hanging -- this happens with multiple engines 86043 960214 build:cgrbeta/75506 - basis_dsizecheck: attempt to write page 66999 on device '/dev/rdsk/c0t2d0s3' failed,system error is:I/O error.DISK INIT encountered an error while attempting to open/create the physical file. ON HP: basis_dsizecheck: attempt 86044 960214 build:/cgrdse/66214: udunmirror: failing over to '/dev/rdsk/c0t3d0s3', i.e. disk mirror mirror = "\$(MIRDEV3)" failed.Failed to mirror device 'master'. This problem is on HP & Solaris. 86373 Solaris's 'intimate shared memory' (ISM) needs to be brought under the 'lock shared memory region' configuration parameter. Currently, there is no way to turn ISM on or off. 86387 The following query will get a seg fault - select name, count(*), count(distinct name) from sysobjects where id = 1 group by all name 86389 Timeslice during renaming current version of config file as part of installmaster. Trace: ... -> cfg_create_file->basis_rename_cfg_file- >basis_get_next_suffix->_fini ... 86537 Customer configures network packet size to either 2K or 4K from 512; after 5 mins. of benchmarking, core dump. Solaris 2.5; SQL Server 11.0. 86538 Customer configures clock tick length to either 20,000K or 120,000K from 100,000; after 5 mins. of benchmarking, core dump. Solaris 2.5; SQL Server 11.0. 86539 select * into tempdb..syslogins from master..syslogins results in a stack trace: sequencer s_compile s_preprocess open_one_side open_range get_range_sdes open_user_table open_byid sdes_hookup dbt_rekeep ex_raise. Works if different table name in tempdb. 86575 Fatal SANITY error 661, EX_TAB lock not held during rollback, in ddlstres test. s_execute->stopxact->abort__all->undo->do_row- >ex_raise(661), Pss->pcurcmd=ABORT. 86600 The error message corresponding to error number MANAGER_BUFINFO has more format parameters in the error message string. The error message takes 2 format parameters (one for bstat and another for bwstat). It caused server to seg.fault on hp platform while Sybase Confidential Page: 4 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 86638 sql server takes 10 seconds to reply to a message sent by the backup server. Occurs when the number of pre-read is filled and bs is waiting for protocol ack. happens on sun4 also. 86656 ct runs sp and gets 702.20.1 Memory request exceeds 3912 the size of single page of 2048 bytes. reproducable at ct site. this is case 10075475. 86793 queries with "exists" subqueries don't always consider all possible join orders, and this sometimes causes a very poor query plan to be chosen. Found in PeopleSoft GL application, where a query runs > 40 minutes, where forcing the "correct" plan allows... 86830 ind_bufguess* counters are being included in two separate lines in the "Task Context Switches Due To" section in sp_sysmon_taskmgmt. Both "Modify Conflicts" and "SYSINDEXES Lookup" include these counters. This causes the total number of known task ...h 86831 sp_sysmon_xactmgmt refers to plc_flush_slr_xact as a System Log Record. actually, this refers to a transaction the contains only a Single Log Record. The output and the documentation needs to be updated. 86897 level 0 scans don't require unique index if table is in read only db. this decision is made at compile time (in q_check_rdonly() called by q_score_index()). if db is changed to writable AFTER the proc is compiled, proc will not be recompile and will... 87037 The use of large buffers for allocation, such as for syslogs in tempdb does not increment buf_grab monitor counters. 87079 cm__cache_wash() keeps a count of pinned buffers that it skips in the variable pincount. This is never used for anything, so either remove it to save the instructions are figure out why it was put in the code in the first place. 87080 cm__cache_wash() calls ucierrlog() to raise the error MANAGER_NAMED_CACHE_TOO_SMAL with the spinlock held. The spinlock should be dropped before writing to the errorlog especially since we are going to drop it immediately to yield the CPU anyway. 87116 [ASID:sp_configure11b/configur_1310/16] expected msg 17417 but server returned 5893 instead 87136 lock_wait() calls the macro MONEVENT_LOCK_WAIT_RESULT after lock__semawait() returns LOCK_DEADLOCK or LOCK_INTERRUPTED. In this case, the macro will SEGV/BUS error because the lock was released. We need to save away the loobjinpageno in a local ... Sybase Confidential Page: 5 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 87280 Monitor counters are incremented while holding spinlocks in various places in the code (e.g., upwakeup()). This causes scaling problems. Also, the MONITOR_INC macro is slow. 87449 In exerbuff test, DIFF in buf001.out: The command "dbcc exer_init_exer(buffmanager) is followed by a message "WARNING!!! Attempt to do a prefetch of size 262144K is not allowed. The prefetch request is reset to the maximum prefetch size of 16384K".e 87476 960304 build:bufman/procs/buf530:sp_unbindcache_all cache3 returns WARNING.You may have to change the log I/O size for the database 'dbase02'.Current setting is '2K' and the cache 'cache3' does not have a '2K' memory pool.This command used to return 40084 1 If you try to kill your own process, you get the message "You cannot use KILL or SYB_TERMINATE to kill your own process." This is bad message because SYB_TERMINATE is undocumented because it leads to corruption. Many customers find out the hard way. 64740 1 There is a regression because of 58532 eco. The file kssndmsg.c should only be built on platforms which support the sendmsg() feature. 69802 1 While running TPC-C benchmark on ruby, the server stacktraced with error 1127.(Internal Error: The OAM has no room for new inserts. Check OAM for object 208003772(orders table), index 0. run->insert- >cinsert->datasplit->insnewpg->linksplit->pg_allocate-> 70023 1 create database occassionaly deadlocks with the checkpoint process. The checkpoint process is chosen as the victim. (Is this okay?) 71428 1 while running regression test sort with MAXOAMENTRY set to 12 (by changing oam.h file #define), error 6902 is detacted in test case sort008.sql. The test was ran with MAXOAMENTRy = 12 to simulate large tables. 72187 1 TEST 0485 SDL032 Priv.violation: SELECT and column UPDATE cursor fail. And also other test cases:0128, 0258, 0261, 0264, 0410, 0413, 0444, 0485, 0489, 0512, 0523 72409 1 Time slice while runningzetatest @_user = 2, @avoid_deadlock =1(err). Stack: kaclkintr->_etext->getnext->xsc__syslogs_getnext- >xls_getnext->def_delete->dodeferred->finish_cmd->s_execute. 73820 1 When the server was booted against a 10.0.2 server, Error 3444 occured. This is "SQL Server could not bring database %s online". The problem is the text is not printed. This could confuse customers. Also is this message expected before running upgrade? Sybase Confidential Page: 6 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 73892 1 Identification of self-joins and self-referential integrity constraints in delete and update queries should be provided, so that queries that do not involve these can be processed in a direct mode resulting in better performance. RI implementation using. 73937 1 In several places, the strmio code unnecessarily holds a spinlock to access engine-private network masks. Cleaning up this will increase concurrency and perhaps improve network I/O performance. 73947 1 950801 build: keyw/kewy0001.sql - alter table u_bin_t add ch12 NCHAR(12) null, vchar NCHAR VARYING(15) null returns: Msg 2705,Level 16, State 2: Column names in each table must be unique. Column 'f' in table 'u_bin_t' is specified more than once. 73974 1 Access method violation: Acquires logical lock during transaction rollback during cursor fixup (getnext) when a row with cursor positioned on it is being deleted as part of rollback and next row is on next page. 75179 1 solaris compiler warnings in these files: access/endxact.c, access/xls_xxs.c, dbcc/dbccdiag.c, dbcc/exec_dbcc.c, demo/conn_hdlr.c, dblkio/duaiobasis.c 76018 1 Two stored procedures are missing: sp_unbindcache_all and sp_helpconfig. Verified on svrhpdly:/svrhpdly_drel1/release/scripts/installmaster of 22 Sept. 76500 1 cm_coalesce and cm_split can potentially create a buffer pool of an invalid size. 76531 1 951002 build - ddlstres/ddls003: 1st run got err msg: Encountered a fatal sever err. Msg 2806,Level 18, state 0, Stored procedure 'ddls_db02..p_ddls_02 is corrupted.Must re-created procedure. Encountered by Spid:13 Kpid:23 User:8. In the 2nd test run 76626 1 This bug is entered to track code changes for the ALLOC part of bug 76486. Page allocations will not be undone correctly after a crash. We must test the performance impact of this change. 76646 1 If a named cache has a 4M 4k pool and if sp_poolconfig "named cache","0","4K" is issued, even if the pool is not completely deleted due to some buffers in use, we should make this buffer pool as not available and any subsequent operation on this should 76852 1 If a table has a update trigger, updates done by the trigger are all being done in deferred mode. Customer feels that these should be done in direct mode to get the desired performance. 77063 1 Table scans are slower in 11.0 than in 10.0.x. One test shows that in-memory scans are 20% slower in 11.0. Sybase Confidential Page: 7 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 77704 1 Multi-user Air Products DSS benchmark performs 158% more IO's when configured with a small 16K buffer pool vs a configuration with all 2K buffers (same total memory for both). Performance does not improve until 16K pool is large. Single user not a problem 78138 1 server stack traced with error 1127, severity 22 state 1 while running the TPC-C benchmark. Stack: run->insert->cinsert->insnewpg- >linksplit->pg_allocate->pg__set_allocbit_oam_loglock->pg__pgalloc- >pg__updateoam->pg__insertoam->ex_raise 78305 1 [forward port of 79622 from 10.0] Executing a replicated sproc after dropping an index on a table involved in the sproc causes the error "Transaction count not zero" after EXECUTE". 78307 1 [forward port of 75736 from 10.0] Logscan thread incorrectly increments curqual for a SAVEXACT log record which is not sent because the field xsptback is -1 (meaning it only marks a SAVE TRAN command, and is not a rollback). 78308 1 [forward port of 39965 from 10.0] A misleading error message is printed when the client network connection cannot write an empty DONE packet to flush the network connection. The message states that '... trailing log transfer info. may be lost.'. 78666 1 Assume we have a 1M of 4K buffer pool in cache foo. If we issue a command "sp_poolconfig "foo","512K","8K","4K"", there is a possibility that 8K pool shall be created with < 512K.( May be more than 128 buffers in the 4K pool are being utilized) . 79248 1 When a user forces data prefetch size to be 16K, the trace shows no 16K pages are fetched. This is probably related to the changes made while fixing 75046. The plan seems to be correct so we might just need to fix the printing routine. 79440 1 Beta-2 refresh:NIST test: Subtest 0506 fails:"SQLSTATE 40001: trans.rollback/serialization fail" Status: fail.Used to pass before update to 1003 connectivity library. 79554 1 951113 build:s2srpc/srpc999: when configure "remote logins" & "remote sites" test returns:'number of remote sites' must not exceed 'number of remote logins' & Verification failed for parameter 'nu 79651 1 After hitting bug 77174, SQL Server shutsdown with signal 11. It was holding Spinlock Resource->rxdesq_spinarray. Stacktrace kisignal>_fini>xxs__pm_markers>plc__lrtopage>plc__flush>xxs__lr_getm arker>xls_getmarker>trig_getloginfo>exectrig>s_execute>... 80077 1 During sybinit upgrade process at each stage we get advancing "..." to indicate the process is in progress. However, following the step where it checkpoints the database and says "done" the process appears to be hung. After 5 minutes we get a message .. Sybase Confidential Page: 8 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 80099 1 disk init gives unclear error message when failing due to a device being too small. 'attempt to write page XXX failed...No such device or address' Customers more likely to see this since we now check for (size - vstart) where we used to just check (size). 80194 1 In Swiss Bank test get "current process infected with 11" followed by Stack trace:crt_work->pg_allocate->pg__set_allocbit_oam_loglock- >pg__locklog_alloc->pg__getpgbuff->bufnewpage- >buf_unhash_and_keep_grabbed_mass->cm_bufunhash->buf_bufprint- >ucierrlog 80197 1 In Swiss Bank test get "current process infected with 10" followed by stack trace:drop_with_retry->dropo->bldind0->logtbldealloc- >pg_objdeall->pg__objdeall_logop->xls_logoperation->plc__flush- >plc__lrtopage->xxs_pm_markers 80221 1 table with image and text fields. slow bcp in is at least twice as fast as fast bcp. no problem on tables without text/image. Some inhouse repor times: 5000 rows in table with int,char(25),text,image 80287 1 The "sp_helpsort" procedure, when executed, has two extra spaces return in the front of the line that show the character sorting order ("! " # $ % & ' ( ) * + , - . / 0 1 2 3 4 5"). 80293 1 Disk init initializes low and high page numbers to be negative when using virtual device numbers larger than 127. The negative numbers do not occur consistently. 80296 1 The sp_helpcache stored procedure will hang when attempting to print out the bind info. Error: 02:95/11/28 16:01:21.95 server Error: 834, Severity: 20, State: 4 02:95/11/28 16:01:21.96 server Illegal attempt to clean buffer: BUF pointer = 0x6aaf38, page 80450 1 ownership of stored procedure is eventually lost/corrupted when called from within another sproc. Results in Error: 2812 " not found...". Occurs on utility sprocs which are called repeatedly. 80496 1 Can configure sortorder to 0, and after that charset to any value (valid or not). When server is rebooted (twice, once in single user mode), the message strings for sortorder and charset are null. 80571 1 Wrong error messages while configuring 'memory alignment boundary'. error msg is correct if configuring less than or more than min./max. limits. Msg 5877:Bad alignment value of 2...valid are powers of 2 in between 512k and 0k inclusive. The units are in b 80608 1 setting the 'memory alignment boundary' parameter to anything other than default causes the server to not boot. Sybase Confidential Page: 9 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 80612 1 cleanup code in stopxact() doesn't always catch all open xdeses. bugid 78921 was added to address the problem of closing an xdes that was opened in openxdes() but never linked in (e.g. not pointed to by pss->pmasterxdes). however, it is possible that... 80615 1 Running sp_poolconfig on busy system beta2 refresh server fails with: 'sddone: Bad address/current process (X) infected with 11/Error at address Y/sddone: read error on virtual disk 2 block Z/kechkfault: engine 1: scheduler infected with signal 11'. e 80696 1 When using clarion 5 raid array without throttle controlled, will fill errlog with sddone: write error on virt disk xx, block yyy, sddone: 2048 bytes passed, 0 returned on write for virt disk xx. Server eventually crashes. cust wants fixed.... 80855 1 Executing an rpc from burlington PDSPS3 server to PUBLIC server at emeryville and vice versa fails with Msg 7211. This has been a intermittent problem. Even after recycling the server the problem 80884 1 2k buffer pool not written to config file. When 16k pool is created 2k pool appears in config file with size=512k, wash=512k 81324 1 Not able to use 'memory alignment boundary' with 4096. After change to 4096, server will not boot, and lots of 'current process (0x0) infected with 11'; 'Address 0x2d4d24 (kechkfault+0x24), siginfo (code, address) = (1, 0xc)' error messages. 81355 1 RPC's between 11.0 servers and 10.0.x and 4.9.2 servers hang with no errors in errorlog. 81413 1 Preupgrade is still using system 10 numbers for some of the checks. All the values need to be reviewed and updated. 81459 1 cm_bufunhash() asserts SPINLOCKHELD, yet in SANITY code, ex_raises (BUFFERM, BWRONGCACHE) without dropping spinlock. This will of course result in shutdown. 81544 1 possibly stress related problem: trying to rpc to 1002/1003 Open Server works fine for several hours, then sql server apparently does not connect to open server anymore. No error message in isql (just print new prompt), dblib client says 'dbresults... 81609 1 A listener with spx/ipx protocol can not be started. 81733 1 On Solaris 2.4, filesystems to which we are dumping seem to delay wr iting the dump file when mounted over NFS. This makes it appear that a dump file does not get written. 81861 1 The eucgb character set is missing from the S-Chinese Language Module. Prevents installation of Simplified Chinese and use of *any* Sybase product in S-Chinese. Sybase Confidential Page: 10 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 82042 1 dbcc printaddress() prints "Invalid input for memory address" even if the address is good. Two problems exist. If is a hex string like "0xc1d5b870", then mpm_hex2long() has a typo that doesn't translate this properly. If is thei 82062 1 after load db on identically segmented db as original (dumped db), followed by alter db with data on, log on, get mixed log and data segment (instead of data only) immediately after load database. 82203 1 Upgrading 4.9 and 10.0 databases to 11.0 does not consider databases that are replicated. Sybinit can not blindly upgrade databases that are being replicated. Procedures for upgrading replicated databases need to be documented and tested. 82473 1 Current process infected with 11 followed by time slice - ucbacktrace->terminate_process->_fini->des_unkeep->closetable- >close_range->des_get->open_lookup->open_by_name->open_user_table- >execproc->copy_tprot->getprocid->des_verify->_coldstar 82541 1 when a unique index is unsafe or scan index is unsafe and we are modifying a variable length column, the update is being done in deferred mode (bug fix 80772). This restriction should be removed 82594 1 960104 buid:logrec/lgrc007a - uselogrec_01 returns Database 'logrec_01' is currently offline; in CANON Database logrec_01 is online. This problem is found on OSF, Solaris, HP & RS6000. 82638 1 SELECT COUNT(unique2F) FROM HPRIME WHERE ... returns "Entering q_score_index() ... Value between step K, K + 1, K = 37" (instead K = 33 in the canon 82649 1 The total # of French, German, and Japanese messages does NOT match total # of English messages. See svrhpdly:/svrhpdly_tst1/tst_sql/960107.rerun/intlmsg for the missing messages. 82665 1 Enhance traceflag 1204 to report the index page number the deadlock is on. Current info is "Process 21 was waiting for a 'address' lock which is useless in determining where the contention is. 82695 1 When using an order by clause on a table that has unique index with order by cols that are in the same order of the index, the server does an extra sort even when this table is an outer member of an outer join involving another table. 82734 1 Certain stored procedures in Swiss Bank beta testing result in a memory request of > 2Kbytes to memalloc. This leads to an ex_raise from memalloc. The memory request should be looked into as the memalloc does not entertain requests > 2Kbytes. Sybase Confidential Page: 11 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 82876 1 Fatal error 8402, state 6, when query concurrent with index drop. ex_raise(8402)pg__get_nextallochint> pg_allocate>pg__allocoam>pg__oamspace>pg__getextents>pg_allocextent> pg__large_allocs>pg_allocate>linkovflow>ncinsert> 83629 1 >= predicate on 1st column of compound key. If only predicate, then optimizer's page estimate is correct. However, if additional predicates on other key columns, optimizer gets wrong page estimate. (We still have to read to end of leaf level since ... 83763 1 execution of the following procedure hangs the server: 'create proc p as set transaction isolation level 0 declare x cursor for select * from sysdatabases set transaction isolation level 3 open x'. problem is probably that we keep recompiling the cursor. 83901 1 A spurious error message received when creating an index in a transaction. This error report causes an application to quit. 84208 1 in tranmgr001.out case while running sp_unbindcache_all on log_cache_1 and on log_cache_2 there are three more extra lines retured -- Message empty. 84291 1 In ISQL, create a nonclustered, composite index with 22 columns on a table that has at least 22 columns. The server crashes with current process infected with 10 AND current process infected with 11. ISQL message: Segmentation Fault - core dumped 84422 1 [ASID:dbcc_bhash2/dbccbuf_0080/3] When dbcc bhash is executed with a cache name, a +ve print_buffers value but a -ve bucket_limit, server doesn't display error msg 2560. 84574 1 sp_unbindcache_all 'log_cache_1' returns "Message empty." versus CANON does not return that string. Found the same diff in bufman/procs/buf530:sp_unbindcache_all caches3.KYT. Sybase Confidential Page: 12 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 84691 1 sp_lockrange @@spid, 0, 0 returned ERROR: PAGE LOCK NOT COUNT WITHIN EXPECTED RANGE; and sp_chkprmt @@spid, 2, 2 got ERROR: PAGELOCKS > High Water Mark and Not Promoted, 84698 1 Filtering is used for OR clause, it should not be. 84707 1 The dbt manager needs to ensure that the contents of a cached DBTABLE are consistent with their corresponding values on disk. SQL Server assumes that the 2 stores are consistent, which may not always be true. 84710 1 SQL Server crashes and core dumps with single line error "current process infected with 11". Occurs when "order by" list is in a different order than the "select" list. 84725 1 Fatal error 661, state 2, no EX lock in backout, in Swiss Bank test, after 1204, out of locks, under open_user_table() (!?). See /remote/qa3/. 84747 1 drop table xxx returns Msg 3712, with message "... or referential integrity contraints" part that is not in CANON's Msg 3712, CANON update required. 84759 1 960201 build:orderby/orby059:negative test - select expression results in more than one column having the same name. Test cathes this error & gives Msg 7348 butthe CANON did not catch the error. 84804 1 Reported counts for deadlocks do not total. sp_sysmon shows 18 total deadlocks, but detail rows show only 9 deadlocks. Output from queries includes 9 'Msg 1205' errors. The '18' shows up in Lock Summary & Avg Deadlocks too. 84851 1 [ASID:union4/union_0230/5,10,20] union4.tcf, cases 8-10 (negative tests). Union statement did not fail for 'real' and 'decimal' in queries involving implicitly convertible datatypes. 84909 1 Segv in ins_sysproc() during swiss bank test. kisignal<_fini. 84971 1 Locks not cleaned up when a sp holding a large number of locks is killed. This also appears if a client has a large number of locks open and a process encounters an error causing a rb and the connection is terminated before rb completes 85094 1 960106 build:bfmgropt/bmo001(prefetch strategy when doing select) return different perfetch strategey.Test returns: generating 10000 rows per scan, using index prefetch wheras CANON returns: generating 10000 rows per scan, using sequential data prefetch. Sybase Confidential Page: 13 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 85354 1 some correlated exists subqueries runs considrably slower with 11 than with 10(sometime hundreds times slower). Small example of behaviour is:select * from sysdatabases d where exists(select * from sysprocedures p ,sysobjects o,sysindexes i ... 85635 1 When have case insensitive dictionary sort order installed, sp_configure with no parameters returns the wrong results. Refer to case 10066464. 85908 1 set isolation level 0 in session - run cursor for update at isolation read committed (isolevel 1) - error 9515 results when updating current row of cursor. Ref manual states that the cursor should operate at isolevel 1 regardless of session opening it.s 86249 1 Three table join that returned the correct two rows in 10.0.2 returns ~19,000 rows in 11.0. When additional SARGs are added the correct rows are returned. 86266 1 When trying to bring database on line Customer sees Error 1127/22/1 followed by internal error and stack pg_insertoam - pg_updateoam - pg_pgalloc - pg_set_allocbit_oam_loglock - pg_allocate - linksplit - indexsplit - srchindex - insert_index_row 86305 1 Message 3121, version 4.9.0 is 259 characters long. The maximum length for any server message is 255 bytes. The maximum length for english server messages is 200 bytes, to allow room for expansion during translation. 86620 1 Same symptoms as bug 69981. Processes hang after issuing "use [database name]" command after one process dies with stack: s_recrtproc-> s_usedb-> dbt_get-> dbt__grab -> dbt__clean -> des_discardall -> des_udesfree. 86647 1 sp that ran fine under System 10 produces a 2805 (Bad pointer ... encountered while remapping) error in system 11 which does not get cleared after dropping and recreating the sp. 86730 1 sp_columns output should flag an identity column as an identity column. table with columns c2 (numeric(10,0) identity) and c3 ( numeric(10,0)). sp_columns display same info for both c2 and c3. script in ~ts/cases/503000s/503225/dseinfo/Repro/script2.sql 86732 1 select count(*) from sysobjects where ((select 1 ) + (select 1 )) between 1 and 10 gives error 204 and stacktrace. replace "between" to (>= and <=) and it works. no problem in system 10 or 492 86785 1 Callers of getnext() should always check for DEADLOCK whenever they access System catalogs as deadlocks are more likely to happen when DDL_IN_TRAN is set in a database, since EX_PAGE locks on system catalogs and user objects would be transactional ones. Sybase Confidential Page: 14 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 87283 1 When an attribute is created such that some OAM entries have to be moved from the first OAM page, the logging of XREC_XOAMENTRYMOVE log record is done incorrectly. It changes contents of first OAM page two times without changing its time stamp. 87311 1 Error 605/21/1 results when we receive an ATTN while trying to lock the allocation page (pg__readalloc()) in the process of dropping the index entries for a worktable. (This work is done as part of closetable() which also drops the work table). Exception. 87402 1 sp_fixindex on many system tables raises "infected with 11" then SQL Server core dumps. Core dump is most likely due to bug# 86731 but at the very least the command SEGV and stack traces. 87509 1 Installing a new SQL Server with a sort order other than the default produces a 2511 error on master..sysconfigures. 87542 1 Using sybinit to change a server with default sort order (50) to any non-default sort order succeeds but causes 2511 errors on sysconfigures. 62334 2 Error message 4814 is incorrect; reports problem with syscolumns then suggests running dbcc on sysobjects. 66993 2 regression in dblkio/dblkio.c caused by changes for bugid 64487, eliminating warning messages through typecasting is incorrect here, checking if an uint variable is less than zero, which can never be true and causes range check to be invalid. 67048 2 If you run installmaster repeatedly, sybsystemprocs..sysprocedures grows until the sybsystemprocs database is full. In 11.0, the limit is reached at 5 installmaster executions, compared to 25 for 10.0.2. 67082 2 sp_setreplicate should allow user with replication role to execute 67087 2 dbcc dbrepair readpage and writepage should check for ts_role 67125 2 Backupserver allows the sp_terminate stored procedure to be called from the server, killing a backupserver thread; this provides the opportunity for users to deny service to other users who need to use backupserver. 67871 2 select @@transtate and select @@trancount should return the status or count of the transaction just before the select is executed. Similar to @@error. 68331 2 [REPSRV] The comments in the header for function lt_init_flush_sync incorrectly state that the most recent checkpoint/commit record is guaranteed to be on disk. Sybase Confidential Page: 15 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 68797 2 kpspinlock is used to guard both run and sleep queue. Having a separate spinlock to guard the sleep queue will reduce the contention on kpspinlock. 68879 2 Client modules of the the Cougar buffer manager should not refer to the bpageno field of a buffer since it may not be initialized. 69423 2 disk init produces device activation error. When a device is dropped using sp_dropdevice followed by a disk init for the same physical device, the device activation error happens. This problem is circumvented by rebooting the server after sp_dropdevice. 69489 2 In the installmaster script, the message number 17490 for sp_dropkey has changed, and is not correct 69908 2 It seems as though the MRU Buffer Replacement Strategy (use and discard) should always be chosen on the last access to a work table as the work table will never be used again. union/unio006.out has 70473 2 join.c attempts to take the # of worktables already in use in the query into account when considering the OR strategy. The calculatio n in j_tables() is incorrect. We need to fix this calculation, and take this info. into account 71248 2 sp_setrepdb should set the replicate status of the database to "true" even if the LTM truncation point is not set. 71859 2 Need to fix old pending bugs 27867 and 40161 related to MIN/MAX optimization in the cougar codeline. 72247 2 syntax sp_configure "user conn", 1000, "default" sets value to 1000. This syntax is ambigous and correct error message should be returned to user -- either use sp_configure "user conn', 0, "default" to set to default or, do not use "default" to set value0 72390 2 Error messages written twice to errorlog. ("verify" option also gets two error messages if options do not pass validation). See conf003e.msg in config test suite outpt, for example. Happens 72392 2 Cannot set 'cnmaxaio_engine' and 'cnmaxaio_server' to their minimum values of 0. Error message does not state reason. Validation fails. If validation is correct, reason should be stated. 72509 2 dbcc help() is not kept upto date with the new dbcc's. E.g., dbcc help (xls). 72732 2 If a table is created with max_rows_per_page feature and then an index is created with fillfactor option. The fillfactor option is not working. Sybase Confidential Page: 16 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 72960 2 Some messages from upgrade look funny; e.g. Adding add sysreferences.frgndbname attribute. 73269 2 Remove config parameter "cmaxdsk" from "cfg_options". Also, remove the old config file in "./startup/cfgfile.c". This cannot be done until bug 73198 is checked in first. 73418 2 In test mbname/mbna002, Msg 102 "Incorrect syntax near 'xxx'" displayed twice, with different values for 'xxx', after 'create' statements with illegal characters. 73442 2 The configuration option names in spt_values do not match the configuration option names to be used with sp_configure. For instance, "total memory" is still "memory" in the table and "additional network memory" is still "additional netmem". 73552 2 attrib_notify() built-in should return "Class cannot be NULL" or "Attribute cannot be NULL" when NULL is passed to the class or attribute parameters. 73853 2 The tli code desperately needs cleanning. There appear to be various logic errors which need to be corrected. The algorithms used in n_tli.c should be synched up with the code in n_socket.c. 73938 2 Preupgrade should give a better estimation of customer's DBs and memory requirement before the upgrade. 74060 2 On an 80 character terminal sp_configure is 'double spaced', ie. an empty line between each config param. This leads to the output scrolling off the screen faster than in previous releases. 74140 2 sp_configure 'option', value where the value is the same as the currently configured value, the configuration code (both sproc and built-in) should return message that nothing was done. 74153 2 Add new stored procedure "sp_helpconfig". 74184 2 This bug is being logged to report various display/formatting problems with the 11.0 sp_configure. These issues are contained in a file on vulcan under the /remote/qa3/bugs/74184 directory. 74312 2 new index output for sp_spaceused provides confusing info for clustered indexes. "size" column is index-pages only, reserved is index + used data pages + unused. 74417 2 If a database is extended beyond 64 GB (causing additional extents to be allocated to the GAM), and then if the database is loaded from a dump that was taken prior to extending the database, then there is a mismatch between the size of the database and th Sybase Confidential Page: 17 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 74679 2 When certain configuration errors occur (and the message is printed through cfg__printerr()) the state does pinpoint the failure point. 74773 2 Query processing code trundles through the plan many times during execution to find out things like (1) last subordinate E_STEP under a given main E_STEP (2) last main E_STEP (3) querytype from last subordinate E_STEP. This happens in execution phase and 74851 2 Monitor counters are not incremented properly when the "lock not needed" condition is encountered in lock_multiple(). lock_logical() increments counters properly, but not lock_multiple(). 74869 2 The units for the config param 'CPU Grace time' should be milli- seconds to match 'time slice' and the documentation. 74941 2 j_cost_formula() contains code, under trace flag 305, which takes values from Pss->pbytespare. Production code should not reference spare fields. To make things worse, trace flag 305 is not documented in trace.h. 75046 2 The jbc_largestmass field was added to the JBUFFER_POOL to remember the largest mass size within the cache for a given object. join.c contains a static prefetchSizes[] with the various mass sizes. This array should be reordered to be [2,16,4,8] and 75157 2 According to PLM func spec, the value configured for 'freelock transfer block size' should never exceed 50% of the max freelocks for an engine. Currently, sp_configure checks this when changing 'max engine freelocks' parameter. However, it doesn't (con't) 75432 2 in cri_cruniq() (cri.c) there is a possibility that all the non- clustered indices will be rebuilt, if we are create a clustered index. we don't store the root pages cached in DES for non-clustered indices and perform proper backout. See cri_main() for a 75509 2 The noobhandler has sections of code hidden using a NOTDEFINED macro. This code should be examined to see if it is worthwhile to use and either kept or removed. Also, need to determine if this functionallity should be moved to the ncheck routine. 75585 2 sp_configure "configuration file" 'restore' and 'write' options write value of 'file descriptors' to the config file written instead of DEFAULT unlike other parameters. (1024 on solaris). Should write out DEFAULT if config value has not been changed. s 75720 2 On all platforms except nt386 and nw386 MEM_EQ() is defined to be memhilo() which does a byte by byte comparison. Could performance be gained by using the system provided memcmp() which is tuned for the platform? Sybase Confidential Page: 18 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 75779 2 It would be helpful if msg 5874 told the user how to correct the problem, e.g. Use sp_configure to increase 'total memory' or decrease configured cache sizes. 75794 2 sp_displaylevel added by bug fix 72638 shows all group names for basic and intermediate levels. Should not show group names whose parameters are not displayed, serves no purpose. May prove to be confusing to users that data is missing. 75986 2 "dump tran master with no_log" by "sa" shows following output: "Attempt by user 0 to dump..." followed by "Attempt by user 1 to dump...was successful". Notice "user 0" in first line becomes "user 1" in second. It should be "user 1" in both the lines. 76172 2 The PLC log space reservation scheme reserves 2 pages per PLC by decrementing the unreserved pages count. If the flush of the PLC causes log page allocation the count is further decremented. This causes the LCT to fire earlier than expected. 76368 2 Offshoot of 75018: The size of virtual range table needs to be only the size of user tables in the statement and not MAXRANGE. The # of user tables is stored in estmt->e_open_varct. This will cut down on size of plan of sproc w/82 SELECT * from sysmessage 76570 2 A view with a sum() aggregate with a group by is created. The query referencing this view has a search clause and an order by. These predicates and order by are not being pushed down the view. 76935 2 "sp_confgiure 'total memory', 2147483647" failed with the error message "The 'total memory', '22000', is not enough to change the parameter 'total memory' to '2147483647'. Run 'sp_helpconfig' to determine the required memory and increase the...y 76976 2 the list of keywords in spt_values (type = "W") does not match sqlkgen. Extra words in spt_values: data_pgs, reserved_pgs, rowcnt, syb_terminate, used_pgs 77110 2 Cannot configure 1000+ users when a 16K buffer pool has been defined in the default data cache. Server says that 2K must be at least 512K, when it is actually 54MB. Getting rid of 16K pool allows one to configure 1000 users 77169 2 sybinit fail to bootup sql server, if the server was not runing when starting sybinit. Error in errorlog file indicates a wrong interfaces file was used. 77173 2 sybinit does not able to verify there is no available device number which we can use as new device for creating 'sybsystemprocs' database. Instead, sybinit will continue upgrade process and fail at task 'create the sybsystemprocs database'. Sybase Confidential Page: 19 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 77202 2 There are a group of error messages documented in the TSG that require the use of dbcc page because they only give minimal information. I have the list of error messages available. 77256 2 In many places of our code we check whether memcreate() returns a NULL PROC_HDR pointer and ex_raise() if this is the case. An example would be in function memalloc(). However function memcreate() will never.... 77397 2 If 'wash size' is missing in the config file for a named cache entry, a misleading error message (wrong format on line nn) is given and the server will not boot. The msg should be 'parameter missing, server will not start' as for other system parameters. 77555 2 Messages "nconnect: t_rcvconnect, An event requires attention" followed by "ninitconn_free: t_unbind, fd = 13, An event requires attention" message seen in the errorlog. May be SQL Server is not hadling the error conditions properly? 77843 2 If wash value is specified in % in config file we should get a proper error message and disallow this option. 77974 2 DBCC DES() need to be modified to accomodate 1) to print objstat.objsysstat2 field 2) Currently it print index cache mapping for all 0-255 indices even if the table doesn't have any index. It should print the mapping only for the indexes that are present 77985 2 min/max online engines error msg is conflicting when setting to illegal value (msg: between 1 and 32) and legal value (msg: between 1 and #of actual cpu's). Either the message should be changed for 2 nd case or original msg has to reflect correct # of cpu 78538 2 The fix for 70943 will not solve the problem of the SARGs for inner tables of a cursor on a join being remapped when the outer table is updated when the cursor is an execute cursor. SARG dependency maps are built when compiling cursor's plan and are 78726 2 Customer pulled controller cable on mirrored device while BCPing. Got 823 error and BCP was killed before the Beta refresh was installed. After refresh installed, no error messages were produced and the bcp "hangs" and doesn't die. 78772 2 [ASID:alter_database2/alterdb_0140/2] "alter database" is invoked in the form "alter database on = n" and "n" is greater than 2048, "alter database" does not fail. 78775 2 [ASID:alter_database2/alterdb_0220/10]When "alter database log on = n" and "n" is greater than 2048, "alter database" does not fail. Sybase Confidential Page: 20 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 78859 2 PG_FIXED bit is not set correctly for the index page when an index is created on an empty table. This causes server to compute offset information unnecessarily on indexes having fixed length columns only. This has a negative impact on performance. 79186 2 dp_configure 'Disk I/O' will not work if the server is configured as case insensitive. It conflicts with 'disk i/o structure'. 79529 2 Alter table set max_rows_per_page fails with error 2715. This syntax is incorrect, and should cause a parser error rather than fail in execution. 79638 2 The multiplication sign is missing in the error message "The sum, (number of user connections) + (number of devices max online engines * 2) +....must be no greater than nnn". Sign is missing between number of devices and max online engines. 79772 2 sp_unbindcache_all fails when dbase & table are bound to a cache. It incorrectly reports "database is in use" 79806 2 While looking in the source code for Error 5858 (symbol RCFG_LOG_CHANGE, I noticed that the message is displayed before we know that we have successfully written the change to the config file. Lance Andersen confirmed. 79837 2 sp_cacheconfig does not display all the cache status. For example a cache can have log only/mixed/default data cache along with HK ignore cache type. It only prints information about mixed/default/log only cache type. It needs to handle the HK ignore h 79972 2 In drop table/drop index a field is set in the descriptor of the object to indicate that a dirty read scan cannot be started on the object. If the drop command fails the field is not reset. So subsequent dirty read scans on the table cannot start up. 79982 2 LOCAL_ENGINE(ep, engine_id) should expand to Engine in case of all platforms. 80101 2 In a xact spanning multiple databases, after the prepare phase, the master-xact is committed. After this stage, if any runtime errors occur, an attempt will be made to rollback the xacn. But if the process is killed after this stage it stays committed. 80102 2 buildmaster -r option was removed (no longer needed with the new external config file) but interactive buildmaster still issues the analogous prompt: 'configuration only?' 80179 2 alignment cost (p_lostcnt) is incorrectly coded. if memory request does not fit on the current page in mem__getspace(), then alignment cost is added to p_lostcnt even though the request is not granted on that page. Sybase Confidential Page: 21 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 80562 2 The error message UTILS_FILEDES_TOO_SMALL during startup says '... + (number of devices max online engines * 2) ...'. when the configured devices, etc would exceed the file descriptor limit. 80754 2 sp_unbindcache_all cache6 got Msg 857: "Command Failed: Cannot bind or unbind database 'dbase01' to or from a cache because it is currently in use" 81163 2 Running two simultaneous sp_poolconfigs on beta-2 refresh server many times results in 'Command Failed: Invalid request to move 144384 buffers from the 2048K pool of the 'Invalid pointer param number 4, pointer value 0xd' cache to the 850797989K pool 81165 2 In message CFG_ERR_CFGOVERWRITE (Warning: The file '%s' has been overwritten) there is a missing / between the directory name and the file name. 81193 2 bug 50430 is a generic problem but was fixed only on non core system 10 platforms , per DSE the fix was not accepted to be ported to generic codeline 10 and 11 because it should be rewritten. 81291 2 sprocs/sprc011.out-- while executing sp_configure "procedure cache er" 80 the Msg 5867 says the washsize for the 2k buffer pool in cache default data cache has been incorrectly configured. It must be a minimum or 10 buffers and a maximum of 80 percent of 81505 2 sp_helpcache returns error 916 in the "Cache Binding Information" part of the output... This only happens when a regular user runs the sp_helpcache command... 82190 2 After creating a cache (sp_poolconfig cache_name, "4M","4k") and then trying to change the entire cache to a different size buffer pool, it fails. If you try to change only a part of the cache to use a different size buffer pool then it works. 82257 2 restart_error() is called twice with the same state, resulting in the same state for two different errors. state 1 in both level0__getnext() and level0_restart(). 82277 2 Tablecount set to 5, using query with subselect of type 'in (select ...)' gives garbage cost (200 digit cost) for subquery plan. all else looks normal for the subquery trace 302. 82846 2 Start two isql sessions in 2 windows. In one, issue a shutdown. Then exit out of the second isql session. The server does not shutdown. You must issue a shutdown with nowait. 82866 2 Error 225 suffers from two duplicate states. The are two instances of state 1, and two instances of state 2. Sybase Confidential Page: 22 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 83348 2 manager/dbcmgr.c:usenextactivedb() assumes that dbt_get()always returns Pss->pdbtable, but it can return NULL. This may cause the keep count on the associated database to be one less than it should be. 83349 2 sequencer/s_constr.c:ri_crrange() calls dbt_get(), assumes it never returns NULL (which it can), and immediately references it as a pointer in the next line of code. This bug could cause a stack trace. 83677 2 [ASID:grant/grant_d9/25] System 11 appears to allow the dumping of device to /dev/null, a violation of statement in REF1 pg 3-179 'You cannot dump to a "null device."' The server merely complaints about /dev/null running out of space. 83732 2 An efficient join plan is not considered by optimizer unless force index is being used. 83754 2 server.loc message # 9101 reads : "This is the correct text after Tech Pubs approval." Can this be the correct entry for this message? 83756 2 server.loc message # 3121 is over the byte limit; it registers as 259 bytes, 4 bytes over the 255 limit. 83757 2 There are a total of 84 server.loc messages that are over the 200 byte limit that is required for localization. These messages need to either be shortened in English, or split up. 84424 2 Create procedure p1, turn off auditing on it with sp_auditsproc, p1,dbname,'off'.Then sp_auditsproc p1 still shows several options enabled e.g: successful deletes, failed deletes, successful updates, failed updates, successful selects, failed selects... 84653 2 dboption "ddl in tran" to true. begin tran. create proc testproc as exec dbo.sp_who. go. exec testproc. Msg2812,level16,state5. Stored proc 'dbo.sp_who' not found.works fine if instead of "dbo.sp_who",we use "exec sp_who or"exec sybsystemproc.dbo.sp_who" 84654 2 Do: 1) create database db1 on dev1 2) alter database db1 log on dev1 with override. Step 2) warns the user that log and data are being mixed on the same disk, when step 2) without override succeeds (correctly). The mixing happened in step 1). 85477 2 sp_sysmon uses group_name = "lock" and field_name = "address_deadlock" to print out "Address" deadlocks. There is no counter by that name, and it seems to repeat the value from the line above ("Shared Page") Sybase Confidential Page: 23 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 85523 2 Upon insert Msg 430 /18/ 1 "Constraint id info in the resvar for check constraints is missing" when column has a rule, of the type like "X%". {X stands for A-Z and 1-9 } 85581 2 In function change_mrowsperpg() we seem to get EX_TAB lock on the target table (user table) just to make sure that it doesn't disappear which we are altering its DES information. Acquiring table level INTENT lock is sufficient. 85636 2 has_subquery() (and sp_procqmode) can't recognize if the sproc contains a subquery when the subquery is proceeded by some other sql statemets. 85676 2 If you specify text index for sp_chgattribute "max_rows_per_page", it fails miserably producing a stacktrace. The stored procedure/underlying function should raise appropriate error message instead of stack tracing. 85727 2 memory can be changed to 2147483647 on AXPOSF. No validation is performed. Test suite config/conf009.out 85729 2 disk i/o structures, max async i/os per engine & server, number of mailboxes, messages and alarms can all be set to maxint values - no validation performed (except on AXPOSF). 85847 2 upgrade 11.0 fails to check that master DEVICE has enough space (3 meg) for post upgrade installmaster to be able to alter master if required. preupgrade needs to have a check for this added 85933 2 User gets error 3307/21/1 and stack trace when issuing a "rollback tran". Multiple 1129 errors follow the 3307 error. 86157 2 dbcc fix_al when correcting 2540 and similar errors should check to see if the page it unlinks for that object is not referenced in sysindexes.[first,last]. If it is referenced then the catalog should also be update. E.g, table has broken page chain with 86397 2 dbcc thresholds, without any parameter stacktrace in the d_callthresholds() routine while dereferencing a NULL pointer (resdom). A check need to be added before deferencing it to verify the proper usage. 86497 2 Select involving right function returns 2 characters at 11.0 where it returned 3 in system 10. Seems to be truncating leading zero. 86501 2 When the 1102 and 3605 trace flags are used to force information about 1105 errors to the errorlog, "No such message" is displayed instead of the text of the 1105 error message. The first line "Error: 1105, Severity: 17, State: 1" IS displayed. Sybase Confidential Page: 24 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 86507 2 When the 1102 and 3605 trace flags are used to force information about 1105 errors to the errorlog, the text "No such message" is displayed rather than the 1105 error message text. The message "Error: 1105, Severity: 17, State: 1" IS displayed. 86558 2 See bug 67744, on disallowing triggers on system tables. Now that it is fixed, warnings against doing this need to be removed from docs (SA Guide and perhaps elsewhere?) Also, this prohibition should be included on the create trigger man page. 86622 2 Pss.pidentity is a part of SEQFRAME and this prohibits retrieving last value of @@identity when inserting to IDENTITY at different nested level of execution. 86849 2 The OS pid for engine 0 is not being printed in the server errorlog, whereas its being printed for all other engines. Reporting this in the errorlog will avoid having to rummage through 'ps -xyz | grep' output. (Need engine 0 OS pid to, for instance, ... 87152 2 Type each of these on their own line: sahutdown, goi, rset, shutdown, go. Process gets infected with signal 11. Stack trace: conn_hdlr->tdsrecv_language->sqlpars->yyparse->etext->_sigtramp- >kisignal->terminate_process->ucbacktrace. 87260 2 Locate an OAM page with space to insert an OAM entry while searching for the existence of an OAM entry before insert 87265 2 'dbcc traceoff' (without any parameters) produces segmentation violation and stacktrace when try to access the null protection description off of the ESTMT in routine exec_dbcc. 87368 2 Error 2812 received when attempting to execute a system stored procedure that calls another stored procedure. Problem does not occur when logged in as sa, but occurs with regular user. Does not occur in 4.9.x nor system 10. 87385 2 Msg 229 : SELECT permission denied on object spt_committab. Connected user = "probe". "probe" user is executing sproc "sp_xa_recover_cs" (part of products XALibrary and XAServer). 87466 2 when creating a table that should take 1962 bytes, error 1701: "creating table 'x' failed failed because row size would be 1964. This exceeds the maximum allowable size of a row in table, 1962." 87549 2 When the user gets 1129 error, it does not report any diagnostic message for tech support or dse to analyse the the root cause of 1129 error. The 1115 trace flag should be used to report some diagnostic message like stack trace, xact id etc. n Sybase Confidential Page: 25 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 57440 3 uaremove() only marks an alarm for deletion, but when set/remove is done in quick succession the server runs out of configured alarms. 57816 3 The trace flag 260 currently turns off doneinproc senddones, this should be changed so that it is turned off by default and turned on with the trace flag. 63204 3 MERGE REVIEW: There is no need to surround BASIS_TERMINATE() in engine.c with #idef EXIT_PROCESSING. The whole point of basis macros is to remove #ifdef's. 64831 3 Server accepts one more connection over what it is configured for. This bug appears in test conn002 64869 3 Info message missing from canon for test suite login, file logn003a.out. Canon needs to be updated. 65374 3 The TOLOWER() macro, used in the typematch() function in parser/typelookup.c is redundant. TO_ASCII() already mapped to lower case. 67085 3 in sp_stored_procedures, "suser_id = 1" should probably be "user_id = 1" 68126 3 memfree() should zero out memory before freeing it when SANITY is defined. This will help find any subsequent references to freed memory by possibly causing a later seg fault. 68357 3 Open this bug for Cougar tracking - duplicate of 38334 69103 3 The diagserver and dataserver executables have the cc command line parameter information inside the string table. This means any customer or competitor can see exactly which files and exactly how we compile each file. This needs investigation. 69176 3 The macro PERIOD is defined in useful/parsetable.c, manager/rvm.c and possibly some other files. There should be a common header file where this is defined and each of these files should include it. 69325 3 Currently in the buffer manager there is no way to specify IOsize for index pages different from that of data pages. So if a prefetch of size 16K is selected for data pages for an index scan, it will do large IOs for index pages and vice-versa. 69414 3 There are many compiler warnings during compilation of SQL Server source files on Solaris 2.3 and 2.4. Some were new with compiler upgrade to version 3.0.1. Compiler warnings should be fixed. 70531 3 SQL Server should return a more meaningful message when creating new login with a name Greater than MAX_LOGIN_NAME_LEN characters. It should return a message such as "Login Name cannot exceed %d characters." Sybase Confidential Page: 26 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 70870 3 sp_helpconstraint scans sysreferences via a cursor and fetch into. A number of the variables being fetched into are not the same time as the column in sysreferences causing a conversion to be done for each column for each fetch into. 71333 3 Sys11 stored procs when run on Sys10 can stack trace. Since there is no procedure upgrade required between 10.x and 11.0 and with the dump compatibility project, it is possible for customers to perform a binary switch AFTER upgrading to Sys11. 71568 3 the clients of qualpage() do not use the appropriate #define for the last parameter to qualpage(). the defines are QUAL_NORM and QUAL_CURS. But, many clients use TRUE, 0, etc. 71845 3 dynamic prepare of a statement containing '... convert(...,?)...' succeeds, paramfmt for the parameter marker is returned as 'TDS_VOID' type. Execution raises error # 206 71883 3 None of the functions in sequencer/open_range.c are defined in the ANSI-C style as they should be. 74054 3 Msg 5851, error message spells 'unkown', should be 'unknown' 76606 3 if the stack size is set quite high say 2147481600. the server boots and starts stack tracing in kpinitstkguard and starts looping in kecheckfault infected with signal 11. 76732 3 set statistics io reports I/O incorrectly for create database. Number of writes is always just slightly > number of pages in db. Create database performs 16K I/O (except for copying model), but statistics I/O reports the results as if it were 2K I/O 76749 3 The file ../manager/sysattrstub.c is obsolete. It should be moved to zOBSOLETE or removed. 76890 3 sdes.srestarts is set in level0_restart(), though it is not referenced anywhere. Since there is already a level 0 restart monitor, this is superfluous. I think it is a vestige of a restart() builtin that was later removed. 77142 3 dbcc findstranded without any arguments stack traced. The stack trace is _coldstart->conn_hdlr->tdsrecv_language->sequencer- >s_execute->exec_dbcc->d_findstranded. 77517 3 Error message 2513 has severity level 16, user error, while the problem is really result of incomplete delete of rows from sysprocedures when user drops a sproc -- Server problem, not user error. Sybase Confidential Page: 27 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 77706 3 sample_sessions - resource.spl invalid attributes: (1) sqlsrv.master_device_name, (2) ...master_device_size: 17, (3) ...default_characterset, (4) ...auditing_device_sysname, (5) ...auditing_device_physname, (6) same as 4 & 5 for sybsystemprocs. 78766 3 [ASID:return3/return_0246/7] When a "return" without an integer expression, is executed in batch, then a return value of 0 is not getting returned 79267 3 When server has max online eng set to 2. 'sp_configure min online engines 2' generates an error message that is confusing and it does not reflect the fact that this option is not being used. 79434 3 [ASID:acos/acos_0050/6]When "arithabort arith_overflow" is set on and a "divide-by-zero" error occurs during a call to "acos" then the Server is supposed to rollback the entire transaction or batch. Rollback does not happen in the case of batch. 79488 3 [ASID:batch_queries2/batchque_0220/5]When "drop database " is used with statements that recreate in the same batch, then the command is supposed to fail with error 2714. The command does not fail, but drops & recreates the database. 79490 3 [ASID:batch_queries2/batchque_0240/7]When "drop index " is used with statements that recreate in the same batch, then the command is supposed to fail with error 2714. The command does not fail, but drops & recreates the Index. 79514 3 [ASID:print3/print_0210/1]When "print" command is called with a format_string that has more place holders than arguments, then the command will fail, but the open transaction is not getting rolled back. 80161 3 [ASID:sp_renamedb1/sp_renamedb_0060/7]When sp_renamedb , is invoked and does not conform to the rule for identifiers, Server is supposed to return message 17240 and set return status to 1. It is returning message 103 instead. 80162 3 [ASID:sp_renamedb2/sp_renamedb_0130:030/6]When sp_renamedb has been successfully invoked, existing views that reference the renamed database is supposed to succeed on execution. But it fails with error 208. 80283 3 [ASID:sp_table_privileges/tablepri_0160:010]When sp_table_privileges is invoked & , having sa_role, has granted a privilege on to , then the "grantor" column in the result should have for rows having as "grantee" Sybase Confidential Page: 28 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 81246 3 Message needs to be cleaned up. Inappropriate for customer to see - Error: 1134, Severity: 20, State: 1 Failed to get lock on OAM page % ld. This should never happen 81340 3 cursor fixup code doesn't seem to handle some errors very gracefully. if an update where current of is rolled back due to full log, it appears that the cursor remains on the current row, so the next fetch returns the same row. 81364 3 select a from table_name prefetch retuns no error. But, select a from table_name,prefetch retuns a 208. Server should retuna an error indicating that the word prefetch is wrong syntex 81947 3 Some platforms, like Sun Solaris, use the setjmp/longjmp flavour whereby the signal masked is saved/restored on each thread-switch. This is an unnecessary overhead. 82048 3 "dump tran to with no_log" completes normally and raises no syntax errors even though the syntax of the command is obviously incorrect. 82196 3 The function ut_handle in crt.c should be handling all errors > EX_INFO and <= EX_USER in the same way, viz., by printing messages and returning EX_BACKOUT. Also, we are returning EX_CONTINUE if the error is an EX_RESOURCE. This needs to be investigated. 83122 3 sybinit does not carry forward user modifications in the RUN_SERVER fileduring upgrade to system11. As an example, the errorlog file location specified by the -e switch. Code should be fixed or docs. need to state this issue. Similar to 68994/48989 83718 3 sp_aux_getsize() comments pertaining to the messages are wrong. The file $DBMS/generic/sproc/aux_getsize has references to a number of messages and these messages are not printed in this stored procedure. A quick check of other stored procedures should... 83988 3 The maximum size of a table row has been fixed to 1962 bytes (for 2Kb SYBPAGESIZEd SQL Servers). This means that for 4Kb SYBPAGESIZEd SQL Servers the row size does not scale up as expected. 85079 3 If you change the first day of the week using DATEFIRST to something other than Sunday the server still counts the weeks in the DATEPART function using Sunday as the first day of the week. When change first day of week should change for all date functions 85693 3 The message for error 8205 reads as follows "Invalid field token passed to descriptor (id = %d)". Most of the places in DES manager code passes 2 arguments (id and field) when raising this error. Either the message or the code need to be altered!!! Sybase Confidential Page: 29 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 86149 3 Help menu for dbcc(page), doesn't list the printoption 3, which is used for printing the control page of a sliced table. 86151 3 ckpt_chores() routine, after setting the database in single-user mode, calls 'kcierrlog' to print the information message. It should use 'ucierrlog'. 86156 3 Unable to put the 'tempdb' in 'single user' mode. The documentation for sp_dboption doesn't say anything specific. I am not sure whether this should be prohibited. In anycase, in dsinit(), we call usedb(..,tempdb,..) but don't call closedb(USEPREV) for it 86240 3 The option "[8] - View/modify page(s)" of the diagnostics menu, causes process to get infacted with 11 or 10. When changing the data at a offset not a multiple of 4 bytes. This is when using trace 4011 and 1606.. 86398 3 dbcc help(thresholds) doesn't print the help information. 86707 3 Reference Supplement, System Tables: sysconfigures page still says that status = 1 means dynamic. But there's a bunch of new bits, the static/dynamic bit seems to be "8" instead. 86904 3 When auto identity is turned on, identity columns are created for tables with primary key constraints and unique keys. This is not the behavior as doc'd or in system 10. 64456 4 dbcc usedextents (0, 3, 1), i.e. with an illegal dbid of 0 causes server to error with 908 complaining about no entries in sysusages. This can happen easily in backup server test suite after a create database has failed. E.g in db01rawc.sql.b 67088 4 files are obsolete and should be removed from the generic/sproc directory: setaudauth.sql, setsa.sql, setsso.sql 68661 4 The buffer header element bpageno should have its name changed to reflect its new meaning. Since bpageno is only valid for the MASS head it should not be used outside the buffer/cache manager code. 71096 4 SCCS string of file sproc/chgattribute is hardcoded. 71319 4 there are way to many similar trace flags in the optimization/join code. 310, 311, 312 and 317 should be considered for consolidation. in particular, 312 is only used for clustered indexes, and 317 is only used for the work plans. 71777 4 looks like trace flag 805 is unused, though it exists in trace.h 72624 4 Forward porting 57273 seems to make 40100 spurious. 40100 should be backed out, to simplify the code. This seems more apropriate to be done in a development codeline than in a production one. Sybase Confidential Page: 30 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 77498 4 This is a dummy bug entry for testing purposes. 79461 4 [ASID:ascii/ascii_0070/7]When the parameter used in "ascii" function is a constant which is not enclosed in single and double quotes, then the expression should fail with an appropriate message. It does not fail for numeric or decimal constants. 79463 4 [ASID:ascii/ascii_0090/9]When "ascii" function is used with a function that returns data which is not directly convertible to char datatype, then the expression should fail. It is not failing for functions which return decimal or numeric data types. 79464 4 [ASID:ascii/ascii_0100/10]When "ascii" function is used with a column datatye which is not directly convertible to char datatype, then the expression should fail. It is not failing for decimal and numeric data types. 79479 4 [ASID:ascii/ascii_0110/11]When "ascii" function is used with a variable datatype which is not directly convertible to char datatype, then the expression should fail. It is not failing for decimal or numeric data types. 79482 4 [ASID:ascii/ascii_0120/12]When "ascii" function is used with a constant which is not directly convertible to char datatype, then the expression should fail. It is not failing for numeric and decimal constants. 79494 4 [ASID:begin_transaction/begintra_0020/2]:When attempt is made to invoke the command "begin transaction label" and the "label" is not a valid identifier, then the Server should fail with error 102. Instead it returns error 103. 79496 4 [ASID:col_name/colname_0022/3]:If the "column_id" parameter to "col_name" is greater than MAX_COL (256), then the function should return NULL. Instead of returning NULL, it returns the column name of colid 1 of the table used for testing. 79498 4 [ASID:compute4/compute_0350/5]:When a "compute" or "compute by" clause using an "order by" clause has a column heading in the "order by" clause, then the statement is supposed to fail with an error. The command does not fail. 79629 4 [ASID:create_procedure4/cproc_0380/2]When an attempt is made to create a procedure that references a table that is not a valid identifier and the "quoted_identifier" is on and the table name is not enclosed in quotes, Server is supposed to give error 102. 79631 4 [ASID:create_procedure6/cproc_0590/7]When an attempt is made to create a procedure with a parameter, and the "default" value specified for that parameter is not constant, Server should fail with error 102.When global variable it does not fail. Sybase Confidential Page: 31 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 79632 4 [ASID:db_id/dbid_0050/5]When db_id function is called with an argument that cannot be implicitly converted to a character datatype, then it supposed to fail. The failure does not happen when 'datetime' datatype is passed to the function. 79633 4 [ASID:disk_mirror2/diskmirr_0280/11]When "disk mirror" is invoked and the specified is already mirrored but with mirror disabled, then "disk mirror" should fail with error 5128.But disk error code 5151 is received instead of 5128. 79634 4 [ASID:index_col2/indexcol_0100/3]When the function "index_col" is called and the datatype of the cannot be implicitly converted to a character datatype, then the command should fail. The function does not fail for decimal & numeric datatype. 79640 4 [ASID:object_id/objectid_0050/5]When the function "object_id" is called with an argument that cannot be implicitly converted to a character datatype, then the function should fail. It does not fail with "datetime" datatype. 79643 4 [ASID:patindex/patindex_0120/12]When "patindex" is called with an function (as 1st or 2nd argument) that does not return data that is implicitly converted to a character datatype, then the function should fail. 79645 4 [ASID:patindex/patindex_0130/13]When "patindex" is called with a column datatype (as 1st argument) that does not return data that is implicitly converted to a character datatype, then the function should fail. 79646 4 [ASID:patindex/patindex_0140/14]When "patindex" is called with a variable datatype (as 1st argument) that does not return data that is implicitly converted to a character datatype, then the function should fail. 79673 4 [ASID:patindex/patindex_0190/19]When "patindex" is called and the third argumnet is used which is not enclosed in quotes and does not begin with using, then error 174 should be generated. Instead error 102 is getting returned. 79677 4 [ASID:print2/print_0160/6]When "print" is called with a format_string that contains invalid place holders arguments, then the command is supposed to fail with error 2754. Instead it fails with error 2747 when '%!' is used as place holder. 79678 4 [ASID:round3/round_0050/1]When arithabort arith_overflow is set on and a divide-by-zero error occurs during a call to "round", then the Server is supposed to rollback the transaction in which the error occured and print a warning message. Sybase Confidential Page: 32 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 79679 4 [ASID:soundex1/soundex_0100/10]When soundex is called with a function that returns data which not directly convertible to char datatype, then the expression is supposed to fail with an error message. It does not fail for decimal, numeric datatype. 79682 4 [ASID:soundex1/soundex_0110/1]When soundex is called with a column d atatype which not directly convertible to char datatype, then the expression is supposed to fail with an error message. It does not fail for decimal, numeric,datetime and smalldatetime. 79684 4 [ASID:str1/str_0150/2]When "str" function is used with 0 arguments, the command is supposed to fail with error message 189. But it fails with error 174. 79686 4 [ASID:str1/str_0120/5]When "str" is called with the third argument as negative it supposed to fail with an appropriate error. The command does not fail when the third argument is a negative. 79687 4 [ASID:str1/str_0110/6]When "str" function is called with the second argument as negative it is supposed to fail with an appropriate error. The command does not fail when the second argument is a negative. 79688 4 [ASID:suser_id/suserid_0040/6]When the function "suser_id" contains more than one argument, then the command should fail with error 174. Instead of error 174, error 189 is getting returned. 79691 4 [ASID:suser_id/suserid_0050/7]When the function "suser_id" is called and cannot be implicitly converted to a character datatype, then the command should fail with an error message. It does fail for decimal, numeric datatypes. 79694 4 [ASID:textvalid/texvali_0060/6]When the datatype of the first argument in a call to "textvalid" function cannot be implicitly converted to a character datatype, the command is supposed to fail. The command does not fail for decimal & datetime datatype. 79696 4 [ASID:textvalid/texvali_0070/7]When the datatype of the "textvalid" function cannot be implicitly converted to binary datatype, then the command should fail with an error message. The function does not fail for character datatype. 79702 4 [ASID:valid_name/validnam_0060/11]When the function "valid_name ()" is called and does not implicitly convert to character datatype, then Serer is supposed to return an error. Error is not returned for smalldatetime & datetime. 79705 4 [ASID:writetext1/writetex_0020/2]When "writetext" is invoked and the datatype of the argument is not varbinary with a length of 16, then the command is supposed to fail. The command does not fail for varchar(30) and nvarchar (30) datatypes. Sybase Confidential Page: 33 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 79709 4 [ASID:writetext3/writetex_0240/10]When"writetext" is invoked on a text column with valid parameters, and the data parameter contains a variable which implicitly converts to a character string, commands succeeds for money & smallmoney but the data is empty 79711 4 [ASID:writetext4/writetex_0245/1]When "writetext" command is invoked on a text column with valid parameters, and the data parameter contains a function which implicitly converts to a character string,the command is supposed to pass. But fail with 186 err. 79712 4 [ASID:writetext4/writetex_0250/2]When "writetext" command is invoked on a text column with valid parameters, and the data parameter contains a variable which does not implicitly convert to a character string,the command does not fail. 79713 4 [ASID:writetext4/writetex_0270/4]When "writetext" command is invoked on a image column with valid parameters, and the data parameter contains a variable which does not implicitly convert to a character string,the command does not fail. 79717 4 [ASID:writetext4/writetex_0280/5]When "writetext" command is invoked on a image column with valid parameters, and the data parameter contains a variable which implicitly convertd to character string, and the 1st 2 chars are not "0x" the comm. should fail. 80113 4 If you do the following you get a row title that gets truncated. That is do "dbcc traceon(3604)", "go", "dbcc memusage" and notice that the row title "default data cache" gets truncated. 80166 4 [ASID:sp_sproc_column2/sproccol_0060:110/11]When "sp_sproc_columns , , , " is invoked, length of the datatype is displayed in one of the columns. The length is returned as zero is case of decimal datatype. 80168 4 [ASID:sp_sproc_columns2/sproccol_0060:140/14]When "sp_sproc_columns , , , " is invoked, and is a datatype created using NULL values, then the 'nullable' column of the result is supposed to return 1. 80173 4 [ASID:sp_table_privileges/tablepri_0220:010/5]When "sp_tables []" is invoked and the server attribute ACCESSIBLE_TABLES is set to "N", then sp_tables is supposed to return tables regardless of whether the user has permission to access them. 80187 4 [ASID:sp_unbindrule/unbindru_0035/6]When "sp_unbindrule" is called with a valid table column & the "futureonly" option, then the Server is supposed to generate an error and set the status to 1. Server does not generate an error. Sybase Confidential Page: 34 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 80196 4 [ASID:sp_table_privileges3/tablepri_0220:010/5]When "sp_table_privileges " is invoked & with "sa_role", has granted REFERENCES to , then the grantor column should contain for all rows with as grantee. 80198 4 [ASID:sp_addlanguage1/addlang_0290/11]When an attempt is made to invoke "sp_addlanguage" & the "shortmonth" parameter contains two or more identical short month names, then the command is supposed to fail. Server does give any error but adds the language. 80200 4 [ASID:sp_addlanguage1/addlang_0300/12]When an attempt is made to invoke "sp_addlanguage" & the "days" parameter contains two or more identical short day names, then the command is supposed to fail. Server does give any error but adds the language. 80201 4 [ASID:sp_addlanguage1/addlang_0230/4]When an attempt is made to invoke "sp_addlanguage" & the current user login has not been granted the "sa_role" then the server is supposed to fail with error 567. Instead it is failing with 229. 80251 4 [ASID:rtrim1/rtrim_0020/2]When rtrim() is called with a function argument which is directly convertible to char datatype & is text of a where clause it will remove trailing blanks in evaluating the where clause. Server is returning error 257. 80267 4 [ASID:sp_addalias/addalias_0030/5]When sp_addalias is invoked with "login_name" which does not exist in syslogins, then the alias is not added, and server is supposed to return error number 17231. The message is getting returned but the msg no. missing. 80272 4 [ASID:exp/exp_0010:010/1]When "exp" function is called and binary value is passed as argumnent, wrong result is getting produced. 80284 4 [ASID:sp_table_privileges2/tablepri_0160:020/10]"sp_table_privileges " is invoked & , having sa_role, has granted privilege on to , then the "grantor" column in the result should contain for rows having as grantee 80304 4 [ASID:create_procedure3/cproc_0260/6]When an attempt is made to use "create view" as part of the body of SQL within a "create procedure" command then the Server is supposed to fail with error 154. Instead error 192 is getting returned. 80309 4 [ASID:create_rule2/crule_0190/5]When an attempt is made to check a value being inserted into a column which has a rule bound to it, and the datatype of the column is incompatible with the datatype used in the rule, then the server should fail with 257. 80387 4 [ASID:delete3/delete_0240/4]Whan a user attempts to delete a row and the user is not the table owner and has not been granted "delete" permission by the table owner, then Server should generate error 262. Instead error 229 is getting returned. Sybase Confidential Page: 35 System Problem Report SYBASE 11.0.1 SQL Server - Mar 10 1996 Sun Solaris SVR4 2.4 SPR No. Pri Description ------- --- ------------------------------------------------------------------- 80391 4 [ASID:create_rule2/crule_0190/5]Whan an attempt is made to check a value being inserted into a column which has a rule bound to it and the datatype of the column is incompatible with the datatype used in the rule then, Server should give error 257. 80394 4 [ASID:create_trigger3/ctrig_0260/6]When an attempt is made to create a trigger and the "SQL" statements contains "create procedure",Server should fail with error 287. But it fails with error 111 instead. 80400 4 [ASID:ifelse/ifelse_0030/3]When an "if" keyword is used without a following logical_expression, then the statement is supposed to fail with error 102. Instead error 156 is returned. 84876 4 in sp_sysmon output, "Total Network I/O Checks:" should not have the colon on the end of the line 84878 4 In sp_sysmon, the output line for "Empty Page Flushes" was supposed to be removed because the counter isn't used. 87373 4 screwy error handling for update in read only db. given x is read only, try "use master; begin tran; update x..x set x = 1". get error 3906 (db is read only) followed by 3901 (this xact uses too many databases -- i.e. exceeded max xdeses).