> <developer@xxxxxxxxxxxx> writes: > >> My database has shutdown several times in the last couple days. I have >> no >> idea why. I am running centos and I have not rebooted the server or >> made >> any configuration changes. I am running postgres 8.2 and it has been >> stable since I installed it about 5 months ago. The databases crashes >> and >> so my software application goes down. When I restart my application >> everything seems to work fine. But then it crashes again, something >> appears to be corrupt. Here are my logs: >> >> >> LOG: server process (PID 501) was terminated by signal 9 > > Signal 9 is SIGKILL which means something outside Postgres is killing > Postgres > processes. Either something is doing kill -9 <pid> of a Postgres pid. > > There used to be some OSes that recorded a SIGKILL process was killed > because > it had run out of memory, but I'm not sure Linux would report it as a > SIGKILL. > What does dmesg say, it doesn't have any OOM messages does it? > > -- > Gregory Stark > EnterpriseDB http://www.enterprisedb.com > > Thanks for the replies... The box is very secure and I think I can safely say no one did a kill -9 on the postgres process. The java application that accesses postgres does sometimes have memory issues but i am surprised this would affect postgres. I am surprised linux allowed one process to affect the other like that. Should i be increasing postgres memory parameters or do you think this might just indicate the box is overloaded? Is there anything i could do logging wise on the postgres side to get a better indication of what is happening? thanks
