Thread: ERROR: xlog flush request not satisfied
I'm seeing the following error in 7.4.6. It first surfaced after remounting a SAN mount from one box to another and then running 'vacuum full': pg_dump: ERROR: xlog flush request 68/7D853080 is not satisfied --- flushed only to 0/15A90A8 CONTEXT: writing block 34 of relation 17156/1470533 pg_dump: SQL command to dump the contents of table "patient_demographics" failed: PQendcopy() failed. pg_dump: Error message from server: ERROR: xlog flush request 68/7D853080 is not satisfied --- flushed only to 0/15A90A8 CONTEXT: writing block 34 of relation 17156/1470533 pg_dump: The command was: COPY public.customer... Here's my pg_controldata: $ pg_controldata pg_control version number: 72 Catalog version number: 200310211 Database cluster state: shutting down pg_control last modified: Mon Mar 6 13:26:19 2006 Current log file ID: 0 Next log file segment: 2 Latest checkpoint location: 0/151E9C8 Prior checkpoint location: 0/14DD760 Latest checkpoint's REDO location: 0/151E9C8 Latest checkpoint's UNDO location: 0/0 Latest checkpoint's StartUpID: 20 Latest checkpoint's NextXID: 2022907 Latest checkpoint's NextOID: 26923 Time of latest checkpoint: Mon Mar 6 13:05:11 2006 Database block size: 8192 Blocks per segment of large relation: 131072 Maximum length of identifiers: 64 Maximum number of function arguments: 32 Date/time type storage: floating-point numbers Maximum length of locale name: 128 LC_COLLATE: C LC_CTYPE: C Here's my pg_xlog directory: -rw------- 1 cudba cu 16777216 Mar 6 08:25 000000680000007F -rw------- 1 cudba cu 16777216 Mar 6 08:51 0000006800000080 -rw------- 1 cudba cu 16777216 Mar 6 09:09 0000006800000081 -rw------- 1 cudba cu 16777216 Mar 6 09:20 0000006800000082 -rw------- 1 cudba cu 16777216 Mar 6 09:30 0000006800000083 -rw------- 1 cudba cu 16777216 Mar 6 09:45 0000006800000084 -rw------- 1 cudba cu 16777216 Mar 6 10:00 0000006800000085 -rw------- 1 cudba cu 16777216 Mar 6 10:10 000000680000007E -rw------- 1 cudba cu 16777216 Mar 6 13:21 0000000000000001 I have pg_filedump built. Any clues on how to tweak this file to restore full access while minimizing data loss? Thanks, Ed
On Monday March 6 2006 12:47 pm, Ed L. wrote:
> I'm seeing the following error in 7.4.6. It first surfaced
> after remounting a SAN mount from one box to another and
> then running 'vacuum full':
>
> pg_dump: ERROR: xlog flush request 68/7D853080 is not
> satisfied --- flushed only to 0/15A90A8 CONTEXT: writing
> block 34 of relation 17156/1470533
> pg_dump: SQL command to dump the contents of table
> "patient_demographics" failed: PQendcopy() failed. pg_dump:
> Error message from server: ERROR: xlog flush request
> 68/7D853080 is not satisfied --- flushed only to 0/15A90A8
> CONTEXT: writing block 34 of relation 17156/1470533
> pg_dump: The command was: COPY public.customer...
^^^^^^^^
Typo: should have been "patient_demographics" as above...
Any clues?
Thanks,
Ed
"Ed L." <pgsql@bluepolka.net> writes:
> I'm seeing the following error in 7.4.6. It first surfaced
> after remounting a SAN mount from one box to another and
> then running 'vacuum full':
> pg_dump: ERROR: xlog flush request 68/7D853080 is not satisfied --- flushed only to 0/15A90A8
> CONTEXT: writing block 34 of relation 17156/1470533
What this looks like to me is corrupt data in a page's LSN field (the
first 8 bytes of the page). The LSN is evidently far larger than it
should be. You might try dumping out the indicated page with
pg_filedump (or even just dd/od) to see if anything obvious jumps out.
regards, tom lane