HELP: Unknown Error Messages From oncheck
Yagnesh
On running oncheck -cI on some of our big tables (~8-10GB) the following
error messages were reported by oncheck. All the indexes that produced such
error message were dropped and recreated. On running oncheck -cI on new
indexes, same error messages were produced.
Any help as to how to fix this will be appreciated.
Following are the details.....
Here's the output from oncheck that was done after the indexes were rebuilt.
--------------------------------------------------------
Fri Jan 31 12:59:52 MST 1997
Validating index pcla_ix_01 for rims:prodadmn.ptchg_ldgr_a...
Index pcla_ix_01
Could not check rowids and perform data<->index check
Unknown error message 0.
Fri Jan 31 14:23:22 MST 1997
Validating index pcla_ix_02 for rims:prodadmn.ptchg_ldgr_a...
Index pcla_ix_02
Could not check rowids and perform data<->index check
Unknown error message 0.
Fri Jan 31 15:24:50 MST 1997
--------------------------------------------------------
At the same time these messages appeared in the online.log file...
--------------------------------------------------------
14:23:22 I/O bad request chunk 3156, pagenum 228477, pagecnt 1
14:23:22 I/O bad request chunk 3156, pagenum 228477, pagecnt 1
14:28:22 Checkpoint Completed: duration was 0 seconds.
14:33:23 Checkpoint Completed: duration was 0 seconds.
14:38:23 Checkpoint Completed: duration was 0 seconds.
14:43:26 Checkpoint Completed: duration was 0 seconds.
14:48:27 Checkpoint Completed: duration was 0 seconds.
14:58:28 Checkpoint Completed: duration was 301 seconds.
15:03:30 Checkpoint Completed: duration was 0 seconds.
15:08:31 Checkpoint Completed: duration was 0 seconds.
15:13:31 Checkpoint Completed: duration was 0 seconds.
15:18:33 Checkpoint Completed: duration was 0 seconds.
15:24:49 Checkpoint Completed: duration was 77 seconds.
15:24:50 I/O bad request chunk 3157, pagenum 64637, pagecnt 1
15:24:50 I/O bad request chunk 3157, pagenum 64637, pagecnt 1
--------------------------------------------------------
Before rebuilding the indexes, the oncheck had produced same error messages
for all the indexes but one. One index produced assret failed message in the
log that and here it is....
--------------------------------------------------------
11:41:53 bfcheck: bad page: pg_flags 4 != type 10
buffer header
c0cf9fc0: c0cfa048 c0cf9f38 00000000 00000000 ...H...8 ........
c0cf9fd0: 00000001 00000000 00000000 00000000 ........ ........
c0cf9fe0 *
c0cf9ff0: 00000000 00000000 00000000 00000000 ........ ........
c0cfa000: 0001638e 07d10001 00006266 5b323534 ..c..... ..bf[254
c0cfa010: 33395d00 00000000 00000000 c0b63a50 39]..... ......:P
c0cfa020: c07a8a18 c0024818 c0be5bc0 00020000 .z....H. ..[.....
c0cfa030: 001e0000 005cad54 c3fb0000 c5da9578 .....\.T .......x
c0cfa040: 00000000 00000080 ........
page
c3fb0000: 005cad54 7676db8a 00000004 001807e4 .\.Tvv.. ........
c3fb0010: 00000000 00000000 ffffffff fffffffe ........ ........
c3fb0020: ffffffff ffffffff ffffffff ffffffff ........ ........
c3fb0030 *
c3fb07f0: ffffffff ffffffff 00000000 7676db8a ........ ....vv..
11:41:53
11:41:53 Assert Failed: Page Check Error in bttreemrg
11:41:53 Who: Session(475, prodadmn@twoface, 24920, 15)
Thread(3414, oncheckm, c5da9578, 1)
11:41:53 Results: Possible inconsistencies in an index of 'rims:"prodadmn".lab
_rslt_a'
11:41:53 Action: Run 'oncheck -cI 5242914'
11:41:53 See Also: /tmp/af.a993b285
--------------------------------------------------------
The oncheck for that index produced this error...
--------------------------------------------------------
Index lab_rslta_idx02
Could not check rowids and perform data<->index check
ISAM error: Unexpected internal error
Index lab_rslta_idx02 is bad. OK to repair it?
Thu Jan 30 11:41:53 MST 1997
--------------------------------------------------------
Hope to find some solution to this...
Thanks.
--
Yagnesh M Mehta | yme...@rencorp.com
Gambro Healthcare | (615) 353-4145
--
Yagnesh M Mehta | yme...@rencorp.com
Gambro Healthcare | (615) 353-4145
David Williams
In article <E4xoy...@harriet.rencorp.com>, Yagnesh
<ren...@harriet.rencorp.com> writes
>
>On running oncheck -cI on some of our big tables (~8-10GB) the following
>error messages were reported by oncheck. All the indexes that produced such
>error message were dropped and recreated. On running oncheck -cI on new
>indexes, same error messages were produced.
>
>Any help as to how to fix this will be appreciated.
>
>Following are the details.....
>
>
>Here's the output from oncheck that was done after the indexes were rebuilt.
>
>--------------------------------------------------------
>Fri Jan 31 12:59:52 MST 1997
>
>Validating index pcla_ix_01 for rims:prodadmn.ptchg_ldgr_a...
> Index pcla_ix_01
>Could not check rowids and perform data<->index check
>Unknown error message 0.
Sounds like either
a) a bug in oncheck
b) something more fundamental is wrong then just index/table
corruption e.g. reserved pages are corrupt, extent information is
corrupt orsystem catalogs are corrupt Try
oncheck -cr
oncheck -ce
oncheck -cc <database name>
>Fri Jan 31 14:23:22 MST 1997
>
>Validating index pcla_ix_02 for rims:prodadmn.ptchg_ldgr_a...
> Index pcla_ix_02
>Could not check rowids and perform data<->index check
>Unknown error message 0.
>Fri Jan 31 15:24:50 MST 1997
>--------------------------------------------------------
>
>At the same time these messages appeared in the online.log file...
>
>--------------------------------------------------------
>14:23:22 I/O bad request chunk 3156, pagenum 228477, pagecnt 1
>14:23:22 I/O bad request chunk 3156, pagenum 228477, pagecnt 1
Again either extent/reserved info is corrupt so online thinks part of
extent is off the end of a chunk or you have hit a bug.
Page Check error in btree merge?. I would run the above oncheck's.
>11:41:53 Who: Session(475, prodadmn@twoface, 24920, 15)
> Thread(3414, oncheckm, c5da9578, 1)
>11:41:53 Results: Possible inconsistencies in an index of 'rims:"prodadmn".lab
>_rslt_a'
>11:41:53 Action: Run 'oncheck -cI 5242914'
>11:41:53 See Also: /tmp/af.a993b285
>--------------------------------------------------------
>
>The oncheck for that index produced this error...
>
>--------------------------------------------------------
> Index lab_rslta_idx02
>Could not check rowids and perform data<->index check
>ISAM error: Unexpected internal error
Again either a bug or corrupt reserved/extent info.
>
>Index lab_rslta_idx02 is bad. OK to repair it?
>Thu Jan 30 11:41:53 MST 1997
>--------------------------------------------------------
>
>Hope to find some solution to this...
>
>Thanks.
>
>--
>Yagnesh M Mehta | yme...@rencorp.com
>Gambro Healthcare | (615) 353-4145
>
--
David Williams
Peter Tashkoff
Check out your physical disk connections,
fibre-optic cables etc, sounds like a hardware
fault.
rgds
Peter T.
}>> David Williams
<d...@smooth1.demon.co.uk> 4/February/1997
09:10am >>>