Thread: posgres-7.1.3 hpux-11.00 posmaster died
Hello, This morning I saw my database had no new record added since yesterday. And my db should be updated on the 15 min with hundreds of records !! I tried first to connect to my db but I got the following answer: >psql tkgdb1240 psql: connectDBStart() -- connect() failed: No such file or directory Is the postmaster running locally and accepting connections on Unix socket '/tmp/.s.PGSQL.5432'? I wanted to restart postmaster with: nohup /opt/postgres/bin/postmaster -i -D /db1240/data 1>>/home/postgres/server.log 2>&1 & but I got this in my server.log: Lock file "/db1240/data/postmaster.pid" already exists. Is another postmaster (pid 6089) running in "/db1240/data"? I displayed the process: >ps -ef | grep post postgres 6089 1 0 Dec 27 pts/ta 0:47 /opt/postgres/bin/postmaster -i -D /db1240/data idxxxxxx 21087 3390 0 16:32:57 pts/ta 0:00 grep post postgres 20590 20580 0 16:20:06 pts/tf 0:00 -ksh idxxxxxx> I started a top but saw nothing special. No hanging process. No vacuum, no psql session, .... and 90 % of free resources. What I see in my server.log is: DEBUG: copy: line 70, XLogWrite: new log file created - consider increasing WAL _FILES DEBUG: MoveOfflineLogs: remove 00000003000000BB But I don't understand what it means :-( Now I will kill the posmaster and start it again. Any help greatly appreciated to understand what's happened. Happy New Years from Belgium. Pierre.
bk226157@skynet.be (pierre) writes: > This morning I saw my database had no new record added since > yesterday. > I tried first to connect to my db but I got the following answer: >> psql tkgdb1240 > psql: connectDBStart() -- connect() failed: No such file or directory > Is the postmaster running locally > and accepting connections on Unix socket '/tmp/.s.PGSQL.5432'? > I displayed the process: >> ps -ef | grep post > postgres 6089 1 0 Dec 27 pts/ta 0:47 > /opt/postgres/bin/postmaster -i -D /db1240/data Looks to me like your postmaster is fine. Probably what is not fine is the socket file /tmp/.s.PGSQL.5432. Do you have some cron job that cleans out /tmp periodically? If so, you might need to teach it not to remove socket files. regards, tom lane
pierre wrote:
Hello,This morning I saw my database had no new record added since
yesterday.
And my db should be updated on the 15 min with hundreds of records !!I tried first to connect to my db but I got the following answer:
>psql tkgdb1240
psql: connectDBStart() -- connect() failed: No such file or directory
Is the postmaster running locally
and accepting connections on Unix socket '/tmp/.s.PGSQL.5432'?I wanted to restart postmaster with:
nohup /opt/postgres/bin/postmaster -i -D /db1240/data
1>>/home/postgres/server.log 2>&1 &but I got this in my server.log:
Lock file "/db1240/data/postmaster.pid" already exists.
Is another postmaster (pid 6089) running in "/db1240/data"?I displayed the process:
>ps -ef | grep post
postgres 6089 1 0 Dec 27 pts/ta 0:47
/opt/postgres/bin/postmaster -i -D /db1240/data
idxxxxxx 21087 3390 0 16:32:57 pts/ta 0:00 grep post
postgres 20590 20580 0 16:20:06 pts/tf 0:00 -ksh
idxxxxxx>I started a top but saw nothing special. No hanging process.
No vacuum, no psql session, .... and 90 % of free resources.What I see in my server.log is:
DEBUG: copy: line 70, XLogWrite: new log file created - consider
increasing WAL
_FILES
DEBUG: MoveOfflineLogs: remove 00000003000000BBBut I don't understand what it means :-(
Now I will kill the posmaster and start it again.
Any help greatly appreciated to understand what's happened.Happy New Years from Belgium.
Pierre.---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
-- ----------------------------------------------------------------------------- Roman Gavrilov Aduva Inc., Web Development Services. work +972-3-7534324 mobile +972-54-834668 romio@aduva.com, romio@netvision.net.il