Re: problems after pg_resetxlog

Lists: pgsql-general
From: Miguel <mmiranda(at)123(dot)com(dot)sv>
To: pgsql general list <pgsql-general(at)postgresql(dot)org>
Subject: problems after pg_resetxlog
Date: 2005-10-28 22:59:22
Message-ID: 4362AD4A.4070705@123.com.sv
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-general

Hi, i had a server crash and after that , postgres refused to start, i
read on the list that pg_resetxlog could help me, indeed, after i run
pg_resetxlog dir_data, postgres started again, but i look a lot of these
errors:

CONTEXT: writing block 1 of relation 1663/17231/69118230
WARNING: could not write block 1 of 1663/17231/69118230
DETAIL: Multiple failures --- write error may be permanent.
ERROR: xlog flush request 98/E3E83848 is not satisfied --- flushed only
to 98/C364344
CONTEXT: writing block 1 of relation 1663/17231/69118230
WARNING: could not write block 1 of 1663/17231/69118230
DETAIL: Multiple failures --- write error may be permanent.
ERROR: xlog flush request 98/E3E83848 is not satisfied --- flushed only
to 98/C368658
CONTEXT: writing block 1 of relation 1663/17231/69118230
WARNING: could not write block 1 of 1663/17231/69118230
DETAIL: Multiple failures --- write error may be permanent.
ERROR: xlog flush request 98/E3E83848 is not satisfied --- flushed only
to 98/C3717F4
CONTEXT: writing block 1 of relation 1663/17231/69118230
WARNING: could not write block 1 of 1663/17231/69118230
DETAIL: Multiple failures --- write error may be permanent.
ERROR: xlog flush request 98/E3E83848 is not satisfied --- flushed only
to 98/C38554C
CONTEXT: writing block 1 of relation 1663/17231/69118230
WARNING: could not write block 1 of 1663/17231/69118230
DETAIL: Multiple failures --- write error may be permanent.

and pg_dump fails with:

pg_dump radius > radius.out
pg_dump: ERROR: xlog flush request 98/D7CFEE64 is not satisfied ---
flushed only to 98/CA37000
CONTEXT: writing block 3683 of relation 1663/17231/69127319
pg_dump: SQL command to dump the contents of table "starttelephony"
failed: PQendcopy() failed.
pg_dump: Error message from server: ERROR: xlog flush request
98/D7CFEE64 is not satisfied --- flushed only to 98/CA37000
CONTEXT: writing block 3683 of relation 1663/17231/69127319
pg_dump: The command was: COPY public.starttelephony (radacctid,
username, realm, nasipaddress, acctstarttime, calledstationid,
callingstationid, acctdelaytime, h323gwid, h323callorigin, h323calltype,
h323setuptime, h323confid) TO stdout;

what can i do?
thanks


From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Miguel <mmiranda(at)123(dot)com(dot)sv>
Cc: pgsql general list <pgsql-general(at)postgresql(dot)org>
Subject: Re: problems after pg_resetxlog
Date: 2005-10-28 23:47:09
Message-ID: 6471.1130543229@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-general

Miguel <mmiranda(at)123(dot)com(dot)sv> writes:
> ERROR: xlog flush request 98/E3E83848 is not satisfied --- flushed only
> to 98/C364344

> what can i do?

Increasing the WAL start address might help (see the pg_resetxlog man
page). Realize that you're probably going to be dealing with corrupted
data, however :-(

regards, tom lane


From: Miguel <mmiranda(at)123(dot)com(dot)sv>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: pgsql general list <pgsql-general(at)postgresql(dot)org>
Subject: Re: problems after pg_resetxlog
Date: 2005-10-28 23:54:18
Message-ID: 4362BA2A.4080901@123.com.sv
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-general

Tom Lane wrote:

>Increasing the WAL start address might help (see the pg_resetxlog man
>page). Realize that you're probably going to be dealing with corrupted
>data, however :-(
>
> regards, tom lane
>
>
>
Yeah, i know, but i only want to be able to finish the dump, i must
execute the

pg_resetxlog again with the -l flag?


From: Miguel <mmiranda(at)123(dot)com(dot)sv>
To: Miguel <mmiranda(at)123(dot)com(dot)sv>
Cc: pgsql general list <pgsql-general(at)postgresql(dot)org>
Subject: Re: problems after pg_resetxlog
Date: 2005-10-29 00:14:21
Message-ID: 4362BEDD.1090602@123.com.sv
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-general

Miguel wrote:

> Tom Lane wrote:
>
>> Increasing the WAL start address might help (see the pg_resetxlog man
>> page). Realize that you're probably going to be dealing with corrupted
>> data, however :-(
>>
>> regards, tom lane
>>
>>
>>
Ops, foget the last email, i reread the man, this is what i have to do:

shiva2 pg_xlog # ls -l
total 180448
-rw------- 1 postgres postgres 16777216 Oct 28 18:05
00000001000000980000001A
-rw------- 1 postgres postgres 16777216 Oct 28 18:07
00000001000000980000001B
-rw------- 1 postgres postgres 16777216 Oct 28 18:10
00000001000000980000001C
-rw------- 1 postgres postgres 16777216 Oct 28 18:10
00000001000000980000001D
-rw------- 1 postgres postgres 16777216 Oct 28 17:25
00000001000000980000001E
-rw------- 1 postgres postgres 16777216 Oct 28 17:28
00000001000000980000001F
-rw------- 1 postgres postgres 16777216 Oct 28 17:37
000000010000009800000020
-rw------- 1 postgres postgres 16777216 Oct 28 17:51
000000010000009800000021
-rw------- 1 postgres postgres 16777216 Oct 28 17:53
000000010000009800000022
-rw------- 1 postgres postgres 16777216 Oct 28 17:46
000000010000009800000023
-rw------- 1 postgres postgres 16777216 Oct 28 17:49
000000010000009800000024
drwx------ 2 postgres postgres 4096 Apr 22 2005 archive_status

So :

pg_resetxlog -l 0x01, 0x98, 0x25 dir_data

Is this ok?