Package: postgresql-8.4 Version: 8.4.12-0squeeze1 Severity: important Every two weeks, this database server stop working because of maximum open files reached by postgres user. This is an excerpt of postgresql.log
2012-09-18 17:16:31 CEST LOG: out of file descriptors: Too many open files in system; release and retry 2012-09-18 17:16:31 CEST STATEMENT: UPDATE agenzia.systemdatabydocument SET creatingUser=$1, creatingDate=$2,lastChangeUser=$3,lastChangeDate=$4,description=$5 WHERE docId=$6 AND seqNr=$7 [...] 2012-09-18 17:16:34 CEST LOG: out of file descriptors: Too many open files in system; release and retry 2012-09-18 17:16:34 CEST STATEMENT: SELECT p.proname,p.oid FROM pg_catalog.pg_proc p, pg_catalog.pg_namespace n WHERE p.pronamespace=n.oid AND n.nspname='pg_catalog' AND ( proname = 'lo_open' or proname = 'lo_close' or proname = 'lo_creat' or proname = 'lo_unlink' or proname = 'lo_lseek' or proname = 'lo_tell' or proname = 'loread' or proname = 'lowrite' or proname = 'lo_truncate' [...] 2012-09-18 17:16:44 CEST LOG: could not accept new connection: Too many open files in system 2012-09-18 17:16:44 CEST LOG: could not accept new connection: Too many open files in system 2012-09-18 17:16:44 CEST LOG: could not accept new connection: Too many open files in system 2012-09-18 17:16:44 CEST LOG: could not accept new connection: Too many open files in system 2012-09-18 17:16:44 CEST LOG: out of file descriptors: Too many open files in system; release and retry 2012-09-18 17:16:44 CEST FATAL: could not open file "global/pg_database": Too many open files in system 2012-09-18 17:16:44 CEST LOG: autovacuum launcher process (PID 12943) exited with exit code 1 2012-09-18 17:16:44 CEST LOG: terminating any other active server processes 2012-09-18 17:16:44 CEST WARNING: terminating connection because of crash of another server process 2012-09-18 17:16:44 CEST DETAIL: The postmaster has commanded this server process to roll back the current transaction and exit, because another server process exited abnormally and po ssibly corrupted shared memory. 2012-09-18 17:16:44 CEST HINT: In a moment you should be able to reconnect to the database and repeat your command. When this happen, it is not possibile to stop the database using the pg_ctlcluster utility: agenzia-f07:~# pg_ctlcluster 8.4 main stop pg_ctl: il server non si sta spegnendo (In English: "server is not stopping"), so I have to manually kill every postgres process before restarting the database using pg_ctlcluster again. When I try to stop it, it write in logfile: 2012-09-19 09:29:34 CEST FATAL: the database system is in recovery mode What does it means "Too many open files in system"? Is it referring to postgresql system or to operating system? I checked /etc/security/limits.conf and found no lines about postgres user. Is this made on purpose? Would changing the maximum number of open files for user postgres change anything? Is this a known problem, at first? Thanks, Giuseppe -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
