Re: streaming replication question

On Sun, Oct 3, 2010 at 1:02 AM, Rajesh Kumar Mallah
<mallah(dot)rajesh(at)gmail(dot)com> wrote:
> As an enduser i setup SR based on
> http://wiki.postgresql.org/wiki/Streaming_Replication
> The master & standby are both powerful machines and are on same gigabit
> switch.
> Things worked as expected I updated 1000's of records in single txn and saw
> them
> appearing on the standby instantly.
>
> Then i wanted  to see the impact of shutting down the standby momentarily .
> This apparently
> failed with following messages in standby logs:
>
> DETAIL:  last completed transaction was at log time 2010-10-02
> 20:57:26.17677+05:30
> LOG:  restartpoint starting: time
> LOG:  received smart shutdown request
> FATAL:  terminating walreceiver process due to administrator command
> LOG:  restartpoint complete: wrote 1568 buffers (0.3%); write=146.237 s,
> sync=0.251 s, total=146.489 s
> LOG:  recovery restart point at 25/EA87BA18
> DETAIL:  last completed transaction was at log time 2010-10-02
> 21:02:21.89303+05:30
> LOG:  shutting down <--- - - - - - - - - - - - - -- - -  shutdown command
> issued here
> LOG:  database system is shut down
> LOG:  database system was shut down in recovery at 2010-10-02 21:02:22 IST
> <---------------- first attempt to start was made
> LOG:  entering standby mode
> LOG:  restored log file "0000000100000025000000EA" from archive
> LOG:  redo starts at 25/EA87BA18
> FATAL:  too many KnownAssignedXids
> CONTEXT:  xlog redo insert: rel 1663/16399/303892827; tid 1503/119
> LOG:  startup process (PID 20527) exited with exit code 1
> LOG:  terminating any other active server processes
>
> LOG:  database system was interrupted while in recovery at log time
> 2010-10-02 20:57:12 IST <--------------- second attempt was made.
> HINT:  If this has occurred more than once some data might be corrupted and
> you might need to choose an earlier recovery target.
> LOG:  entering standby mode
> LOG:  restored log file "0000000100000025000000EA" from archive
> LOG:  redo starts at 25/EA87BA18
> FATAL:  too many KnownAssignedXids
> CONTEXT:  xlog redo insert: rel 1663/16399/303892827; tid 1503/119
> LOG:  startup process (PID 20693) exited with exit code 1
> LOG:  terminating any other active server processes

This looks like the bug of HS. But I'm not sure why KnownAssignedXids
overflowed.

Regards,

--
Fujii Masao
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center