Sergey Konoplev wrote:
> I'm sorry I mean not HUP but KILL
Hmm...
>> datname | usename | procpid | current_query | waiting |
>> query_start
>> -----------+------------------+---------+-----------------+---------+-------------------------------
>> transport | belostotskaya_la | 20530 | select * from c | f |
>> 2007-10-02 05:05:28.908687+04
>> transport=# select pg_catalog.pg_cancel_backend(20530);
>> pg_cancel_backend
>> -------------------
>> t
>> datname | usename | procpid | current_query | waiting |
>> query_start
>> -----------+------------------+---------+-----------------+---------+-------------------------------
>> transport | belostotskaya_la | 20530 | select * from c | f |
>> 2007-10-02 05:05:28.908687+04
1. Is it always the same query?
2. Does the client still think it's connected?
3. Is that query using up CPU, or just idling?
4. Anything odd in pg_locks for the problem pid?
--
Richard Huxton
Archonet Ltd