In our <SID>_BS.log file, we've received the following errors during a database dump with verify headers.
Apr 1 04:40:01 2014: D01: Page Verification Error: Object id wrong: alloc page=140226816 extent id=140226848 page#=14022685
5 ptnid in ext=8 ptnid in page=-1.
Apr 1 04:40:01 2014: D01: SYBMULTBUF WARNING: One or more errors have been reported while performing consistency checks on the pages being dumped. The affected pages will be reread in an attempt to discard spurious errors.
Apr 1 04:40:01 2014: D01: SYBMULTBUF WARNING: After rereading the pages, the consistency check has continued to report one or more errors.
May 10 18:05:53 2014: D06: Page Verification Error: Object id wrong: alloc page=138916608 extent id=138916712 page#=13891671
7 ptnid in ext=8 ptnid in page=-1.
May 10 18:05:53 2014: D06: SYBMULTBUF WARNING: One or more errors have been reported while performing consistency checks on the pages being dumped. The affected pages will be reread in an attempt to discard spurious errors.
May 10 18:05:53 2014: D06: SYBMULTBUF WARNING: After rereading the pages, the consistency check has continued to report one or more errors.
Looking into what we can do to check for consistency problems, it appears most people were recommending dbcc checkstorage. This requires its own special database, which is planned out with the sp_plan_dbccdb below. Looking at this, we can see it's going to require a lot of space and cache for the <SID> database. The database is > 7 TB compressed. Also because of its size, alternatives such as checkdb and checkalloc will take a TON of time (according to sapnote 1990528: 960GB = ~ 4+ days).
Recommended size for dbccdb database is 157614MB (data = 157612MB, log = 2MB).
No suitable devices for dbccdb in master..sysdevices.
Recommended values for workspace size, cache size and process count are:
dbname scan ws text ws cache comp mem process count
master 21504K 5504K 5402K 0K 1
tempdb 1968M 492M 492M 0K 2
model 1408K 384K 640K 0K 1
sybsystemdb 3712K 1024K 1280K 0K 2
sybsystemprocs 25216K 6400K 6324K 0K 1
sybmgmtdb 12672K 3200K 3175K 0K 1
<SID> 94G 24G 24G 0K 77
saptools 3318M 830M 830M 0K 2
saptempdb 9585M 2397M 2397M 0K 1
So my questions: Is dbcc the best way to troubleshoot the above errors in the backup server log (any other ways you would recommend)? Does dbccdb really need that amount of space and cache size for the <SID> database and what is the minimum it would need? It would be awesome to run a dbcc checktable against the tables we know are corrupted, but finding which tables are corrupt is the issue (lack of time and resources).
Side question: We've had checkalloc running since 5/12, did we have to spool it to an output file or will it generate something at the end? (If we ever get to the end, but regardless, we can't wait weeks to find corrupted tables/indexes)
Currently ASE 15.7 SP110 on AIX