On Friday, December 28, 2012 10:58 AM Harry wrote:
> Hi,
> Kindly see the below attached statements related to Cursor Fetch Issue it's
> still residing as a process.
>500 20222 31036 79 Dec27 ? 16:22:31 postgres: user1 sampledb 192.168.0.40[36022] FETCH
>500 20829 31036 81 Dec27 ? 16:18:48 postgres: user1 sampledb 192.168.0.40[57591] FETCH
>500 20867 31036 81 Dec27 ? 16:09:33 postgres: user1 sampledb 192.168.0.40[45316] FETCH
>500 20870 31036 81 Dec27 ? 16:09:12 postgres: user1 sampledb 192.168.0.40[45343] FETCH
Not sure if FETCH has hanged due to some reason or some other problem due to which cancel or terminate backend is not
working.
Can you provide a stacktrace of hanged backends? Also do let me know you Postgresql version and OS?
Just for test, can you connect a new backend with psql. verify the entry for same in pg_stat_activity. call terminate
APIfor this session. check again if the new entry has gone from pg_stat_activity?
With Regards,
Amit Kapila.
