On Saturday 17 September 2005 18:25, Jose Ildefonso Camargo Tolosa wrote: > Hi! > > I'm about to do some tests with PostgreSQL backend, I just can't keep > the berkley work find. If you can't configure the most commonly used backend to work reliably ... > This is my DB_CONFIG , and it just keeps > crashing, with very low load and a small directory: > > set_cachesize 0 33554432 1 > set_lg_regionmax 262144 > set_lg_bsize 2097152 > #set_lg_dir /var/log/bdb > # > # Automatically remove log files that are no longer needed. > set_flags DB_LOG_AUTOREMOVE > # Just use these settings when doing slapadd... > #set_flags DB_TXN_NOSYNC > #set_flags DB_TXN_NOT_DURABLE > > So, I decided to give a try to postgres backend, and I would like to > hear any experiencies you may have with it. > > I'm usign db-4.2.52, and openldap 2.2.28. With very little load, at > one moment the slapd just go to 99% processor use (it still works, but > it happend), and then I just kill it (signal 9, signal 15 doesn't kill > it) and run db4.2_recover and it starts working just fine for another > day or so. This far, I have not had any data loss, but it is annoying > to see 100% CPU using all the day. How many entries do you have in your database (DB_CONFIG entries aren't that useful without an idea of your database size)? Have you looked at the output of db_stat -m? Are all attributes used in search filters indexed? Anyway, my experience with the bdb backend is that it has been stable ... but you might have other circumstances (so, you may want to give more information). Regards, Buchan -- Buchan Milne ISP Systems Specialist B.Eng,RHCE(803004789010797),LPIC-2(LPI000074592)
Attachment:
pgppsYJdIhyee.pgp
Description: PGP signature