[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
RE: OpenLDAP as an enterprise level LDAP provider
We worked around this by running db_checkpoint in the backround and
checkpointing every 200 Meg of log entries. I didn't know about
DB_TXN_NOT_DURABLE until Quanah mentioned it earlier. Also, as many times I
looked at the DB_CONFIG, I though I was at 520 Meg, not 52 Meg. Duhhh.
After making these changes, and uping the cache to 1 Gig, the slapadd, with
indexes, ran for 65 minutes. Much better!!
Thanks!!
John Fortin
PBG Middleware and Web Services
(914) 767-7844
>-----Original Message-----
>From: Howard Chu [mailto:hyc@symas.com]
>Sent: Wednesday, November 24, 2004 9:50 PM
>To: Fortin, John {PBG}
>Cc: OpenLDAP Mail List
>Subject: Re: OpenLDAP as an enterprise level LDAP provider
>
>
>Fortin, John {PBG} wrote:
>
>>Thanks Howard. I'll try a new build, but I have done it from
>scratch twice
>>and actually looked at the code to ensure the patches were
>applied. I'll
>>also try a build with bdb4.3 and 2.2.18 with Quanah's patches.
>>
>>
>>
>It just occurred to me what you're actually doing. (ding...)
>The slapadd
>tool doesn't do any checkpointing while it runs, so DB_LOG_AUTOREMOVE
>doesn't have any opportunity to work. It never seemed to be important
>before because we always disable transaction logs when running
>slapadd.
>If you (using BDB 4.2) add "set_flags DB_TXN_NOT_DURABLE" to your
>DB_CONFIG file (and you increase your cache from your current 52MB to
>something larger) you should be able to slapadd your 600K entries
>without any problem. Just remember to comment out the flag before
>starting slapd.
>
>In CVS HEAD / release 2.3 we've added a "fasttool" slapd.conf option
>which automatically sets this flag for you when running slapadd /
>slapindex, so you don't need to remember to muck with the
>DB_CONFIG file.
>
>--
> -- Howard Chu
> Chief Architect, Symas Corp. Director, Highland Sun
> http://www.symas.com http://highlandsun.com/hyc
> Symas: Premier OpenSource Development and Support
>