Thread: Runaway postgres process?
Hi. I just discovered three "postmaster" processes running and really eating all the cpu. I'm not sure why? So, what steps do I take for finiding the queries that is causing this?
Well, it stoped. And in the log I can see: LOGG: statement: UPDATE VList SET updated=true WHERE id=985 LOGG: duration: 1129993.209 ms and a few lines below (with queries in between) LOGG: statement: UPDATE VList SET updated=false WHERE id=985 LOGG: duration: 0.400 ms So, something was blocking the query, seriously! Still interested in findingout how to detect what is going on. There were other processes, could it be a lock? I have not specified any special transaction or so. I was calling from a PHP script (I think). 2009/1/30 A B <gentosaker@gmail.com>: > Hi. > I just discovered three "postmaster" processes running and really > eating all the cpu. > I'm not sure why? > So, what steps do I take for finiding the queries that is causing this? >
2009/1/30 Laurent Wandrebeck <l.wandrebeck@gmail.com>: > 2009/1/30 A B <gentosaker@gmail.com>: >> Hi. > Hello, >> I just discovered three "postmaster" processes running and really >> eating all the cpu. >> I'm not sure why? >> So, what steps do I take for finiding the queries that is causing this? > select * from pg_catalog.pg_stat_activity; Ah, great! Some things: In the field query_start it is empty, and the current_query field says: "<command string not enabled>" but that seems to be fixed by turning on "stats_command_string". What does the backend_start field indicate?