Re: [BUGS] deadlocks in postgresql 7.2.1

2003-07-28 Thread Dmitry Tkach
Thanks a lot for the reply! The applications uses the jdbc driver with autocommit turned off, commit and rollback. Do you know if the jdbc driver just starts a new transaction as soon as the last one was ended with commit/ rollback ? Yes, it does. You have to explictly setAutocommit(true) in

Re: [BUGS] deadlocks in postgresql 7.2.1

2003-07-28 Thread Tom Lane
Philipp Reisner <[EMAIL PROTECTED]> writes: >762 ?S 0:00 /usr/lib/postgresql/bin/postmaster >764 ?S 0:00 postgres: stats buffer process >765 ?S 0:00 postgres: stats collector process > 24872 ?S 0:00 postgres: sd sd 10.2.2.6 idle in tr

Re: [BUGS] deadlocks in postgresql 7.2.1

2003-07-28 Thread Philipp Reisner
Am Montag, 28. Juli 2003 11:41 schrieb Peter Eisentraut: > Philipp Reisner writes: > > Once in a while (about 3 times a day) one or more INSERTS/DELETES simply > > go into the "waiting" state, and block the whole database. The only way > > out is to terminate the client connection (i.e. to abort th

Re: [BUGS] deadlocks in postgresql 7.2.1

2003-07-28 Thread Peter Eisentraut
Philipp Reisner writes: > Once in a while (about 3 times a day) one or more INSERTS/DELETES simply > go into the "waiting" state, and block the whole database. The only way > out is to terminate the client connection (i.e. to abort the blocked > INSERT/DELETE query) > > Further investigation with

[BUGS] deadlocks in postgresql 7.2.1

2003-07-28 Thread Philipp Reisner
Hi, We use postgresql 7.2.1 (actually the Debian binary packet version 7.2.1-2woody2) Our database load has more QUERIES than INSERT/DELETES. But we have a sever problem with the wirting operations. Once in a while (about 3 times a day) one or more INSERTS/DELETES simply go into the "waiting" s