--On Sunday, April 15, 2007 11:22 AM +0200 Gyuris Szabolcs
<[EMAIL PROTECTED]> wrote:
Hi Quanah!
Here it is the slapd.conf.
But I think the large number of threads is because after the slapd froze
then all connected client froze too. :(
Hi Gyuris,
In looking at your slapd.conf file, I don't see any checkpoint directive.
When slapd goes to shutdown, one of the things it does is flush everything
from the Berkeley DB cache to disk. If you do not have a checkpoint
directive, and a large number of writes have occurred, this can be
particularly time consuming, as all changes since the last startup must be
flushed. Setting the checkpoint directive will have slapd periodically
write out what changes have occurred to disk, so that the amount of time it
takes to shut down is significantly reduced.
From the slapd-bdb(5) man page:
checkpoint <kbyte> <min>
Specify the frequency for checkpointing the database
transaction log. A checkpoint operation flushes the
database buffers to disk and writes a checkpoint record
in the log. The checkpoint will occur if either
<kbyte> data has been written or <min> minutes have
passed since the last checkpoint. Both arguments
default to zero, in which case they are ignored. When
the <min> argument is non-zero, an internal task will
run every <min> minutes to perform the checkpoint. See
the Berkeley DB reference guide for more details.
I'm not 100% sure that this is the issue you are having, but it is
certainly a possibility. Can you try adding the directive and see if it
helps? I've been using back-hdb on my directory deployment @ Stanford for
the entire 2.3 release and haven't had issues with it shutting down
properly. It is hard to tell from the information here if the issue is
really slapd locking up, or simply slapd taking a while to shut down.
Regards,
Quanah
--
Quanah Gibson-Mount
Senior Systems Software Developer
ITS/Shared Application Services
Stanford University
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
