Thread: please help--database just locks up.
we are running linux and postgres7.3.2 in a production environment.
the database just 'locks ups.' if i try to get into psql X [where X is my database or template 0 or template 1] it just sits there.
the log files don't say anything......there's no indication as to what is happening.
if i look at top it says 100% idle, and the postmaster is supposedly still running, as i see it when i do a ps auuwx| grep ^postgres....
can anybody please help me figure out how to find out what's happening?
thank you so much,
Mona Gamboa
ok...here's the output.[kinda nasty to read] uname -a Linux aus-db-1.loopone.com 2.4.9-e.3enterprise #1 SMP Fri May 3 16:35:33 EDT 2002 i686 unknown [postgres@aus-db-1 postgres]$ grep ^log_ $PGDATA/postgresql.conf log_connections = true #changed this whole block to true -- by Mona G. log_pid = true log_statement = true log_duration = true log_timestamp = true log_min_error_statement = error # Values in order of increasing severity: [postgres@aus-db-1 postgres]$ cat $PGDATA/postmaster.pid 2411 /home/db-pri/postgres/pgsql/DBTRANS 5432001 98306 [postgres@aus-db-1 postgres]$ ps augxwww | egrep 'postgres|postmaster|PID' USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND postgres 2398 0.0 0.9 211248 14200 ? S 02:38 0:29 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2407 0.0 0.9 211248 14200 ? S 02:38 0:00 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2408 0.0 0.9 211248 14200 ? S 02:38 0:01 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2409 0.0 0.9 211248 14200 ? S 02:38 0:00 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2410 0.0 0.9 211248 14200 ? S 02:38 0:00 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2411 0.0 0.1 24380 2256 ? S 02:38 0:00 /home/db-pri/PostgreSQL/bin/postmaster postgres 2412 0.0 0.9 211248 14200 ? S 02:38 0:00 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2413 0.0 0.9 211248 14200 ? S 02:38 0:00 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2414 0.0 0.9 211248 14200 ? S 02:38 0:00 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2415 0.0 0.9 211248 14200 ? S 02:38 0:00 /usr/local/jdk1.4.1/bin/java -Dlog.base.name=/home/db-pri/postgres/logs/db-p ri_dbtrans -Dsend.email=false -Dsave.errors=true -Dsave.queries=false -Dmess age.threshold=300 com.loopone.logrotator.LogRotator postgres 2416 0.0 0.2 25336 3216 ? S 02:38 0:01 postgres: stats buffer process postgres 2417 0.0 0.1 24508 2352 ? S 02:38 0:01 postgres: stats collector process postgres 11095 0.0 0.0 2508 1332 pts/1 S 13:46 0:00 -bash postgres 13028 0.0 0.1 2856 1712 pts/1 S 15:35 0:00 ssh aus-db-1 postgres 13032 0.0 0.0 2516 1340 pts/4 S 15:35 0:00 -bash postgres 13964 0.0 0.2 25144 4108 ? S 16:24 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 13969 0.0 0.2 25144 4104 ? S 16:24 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 13972 0.0 0.2 25144 4104 ? S 16:24 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 13974 0.0 0.2 25124 4092 ? S 16:24 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 13984 0.0 0.2 25144 4112 ? S 16:24 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 14069 0.0 0.2 25124 4092 ? S 16:28 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 14070 0.0 0.2 25124 4092 ? S 16:28 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 14072 0.0 0.2 25124 4092 ? S 16:29 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 14073 0.0 0.2 25124 4092 ? S 16:29 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 14077 0.0 0.2 25124 4092 ? S 16:29 0:00 postgres: postgres dbtrans 10.1.1.130 idle postgres 14478 0.0 0.0 2520 1336 pts/3 S 17:08 0:00 -bash postgres 14665 0.0 0.0 2508 1328 pts/2 S 17:23 0:00 -bash postgres 14717 0.0 0.0 2844 912 pts/2 R 17:24 0:00 ps augxwww postgres 14718 0.0 0.0 1724 596 pts/2 S 17:24 0:00 egrep postgres|postmaster|PID [postgres@aus-db-1 postgres]$ psql -d template1 -c "select version()" version ------------------------------------------------------------- PostgreSQL 7.3.2 on i686-pc-linux-gnu, compiled by GCC 2.96 (1 row) [postgres@aus-db-1 postgres]$ uptime 5:25pm up 14:52, 4 users, load average: 0.02, 0.03, 0.00 [postgres@aus-db-1 postgres]$ df $PGDATA Filesystem 1k-blocks Used Available Use% Mounted on /dev/sdb6 15124868 2930408 11426156 21% /home/db-pri/postgres thanks, Mona ----- Original Message ----- From: "Ed L." <pggeneral@bluepolka.net> To: "Mona Gamboa" <mona@loopone.com> Sent: Monday, February 17, 2003 4:00 PM Subject: Re: [ADMIN] please help--database just locks up. Can you post the output of these commands? uname -a export PGDATA=/path/to/your/data grep ^log_ $PGDATA/postgresql.conf cat $PGDATA/postmaster.pid ps augxwww | egrep 'postgres|postmaster|PID' psql -d template1 -c "select version()" uptime df $PGDATA Regards, Ed On Monday February 17 2003 3:16, you wrote: > we are running linux and postgres7.3.2 in a production environment. > > the database just 'locks ups.' if i try to get into psql X [where > X is my database or template 0 or template 1] it just sits there. > the log files don't say anything......there's no indication as to > what is happening. > > if i look at top it says 100% idle, and the postmaster is > supposedly still running, as i see it when i do a ps auuwx| grep > ^postgres.... > > can anybody please help me figure out how to find out what's > happening? > > thank you so much, > > Mona Gamboa
On Monday February 17 2003 4:31, Mona Gamboa wrote: > ok...here's the output.[kinda nasty to read] Nothing obvious to me. If it is reproduceable, you might try turning on some of the lock tracing directives in postgresql.conf and/or increasing the logging level. You might also try attaching to the hung process in gdb to see where it is hanging, and post the backtrace and see if anyone recognizes anything, e.g. gdb <PATH-TO-PSQL> <PID> gdb> bt Ed
Can you post the output of these commands? uname -a export PGDATA=/path/to/your/data grep ^log_ $PGDATA/postgresql.conf cat $PGDATA/postmaster.pid ps augxwww | egrep 'postgres|postmaster|PID' psql -d template1 -c "select version()" uptime df $PGDATA Regards, Ed On Monday February 17 2003 3:16, Mona Gamboa wrote: > we are running linux and postgres7.3.2 in a production environment. > > the database just 'locks ups.' if i try to get into psql X [where > X is my database or template 0 or template 1] it just sits there. > the log files don't say anything......there's no indication as to > what is happening. > > if i look at top it says 100% idle, and the postmaster is > supposedly still running, as i see it when i do a ps auuwx| grep > ^postgres.... > > can anybody please help me figure out how to find out what's > happening? > > thank you so much, > > Mona Gamboa
Is it reproduceable after a restart? Can you dump the production data, reload on a test cluster, and reproduce it with any consistency? Can you post the output of these commands? uname -a export PGDATA=/path/to/your/data egrep '^log_|lock' $PGDATA/postgresql.conf cat $PGDATA/postmaster.pid ps augxwww | egrep 'postgres|postmaster|PID' psql -d template1 -c "select version()" uptime df $PGDATA Regards, Ed On Monday February 17 2003 3:16, Mona Gamboa wrote: > we are running linux and postgres7.3.2 in a production environment. > > the database just 'locks ups.' if i try to get into psql X [where > X is my database or template 0 or template 1] it just sits there. > the log files don't say anything......there's no indication as to > what is happening. > > if i look at top it says 100% idle, and the postmaster is > supposedly still running, as i see it when i do a ps auuwx| grep > ^postgres.... > > can anybody please help me figure out how to find out what's > happening? > > thank you so much, > > Mona Gamboa