--On Thursday, August 16, 2007 3:38 PM -0700 Sumith Narayanan sumith.narayanan@gmail.com wrote:
Hi All,
Thanks for the suggestions :
I will look forward to upgrade the systems to 64 bit processors and increase the memory to 16 GB, since it will take time , I am doing some more tweaking to see whether it will help.
I reduced the set_cachesize in the DB_CONFIG by 10 times and also in the slapd.conf file. Restarted the process and it was consuming less memory now. But the master from whether the updates flow to slave is still slowing using up the memory and crashing , but the slave seems to be stable with a usage of around 200 MB of memory. I don't think I will be able to avoid that with the current situation.
I have few questions now :
- Any idea on how big the openldap DB can be ?
- I see some DB transaction files in the db directory , which are of
same size and more than one is created in a day :
-rw------- 1 ldapp ldapp 20485760 Aug 13 21:40 log.0000001070 -rw------- 1 ldapp ldapp 20485760 Aug 14 05:52 log.0000001071 -rw------- 1 ldapp ldapp 20485760 Aug 14 16:08 log.0000001072 -rw------- 1 ldapp ldapp 20485760 Aug 14 23:06 log.0000001073
I took a backup of a lot of this inorder for the slapd process to start faster. It keeps on getting accumulated. I guess these are DB transaction log files. Is it supposed to get deleted automatically ? Why is it accumulating ?
Go read the FAQ entry on tuning BDB. If you don't set the flag to have them automatically remove when checkpointed, then they'll stay forever. Up to you whether or not you want to keep them, there are valid reasons for doing so.
--Quanah
--
Quanah Gibson-Mount Principal Software Engineer Zimbra, Inc -------------------- Zimbra :: the leader in open source messaging and collaboration