On Sat, Jun 18, 2005 at 10:52:31PM +0200, Patrick Willam wrote: > Geert Stappers wrote: > > > >Could you tell more about the performance of the server? > > > >I think the machine is too busy when nagios is checking ldap. > > It's an Athlon >1.5GHz with 2GB RAM.
Mmm, that is surely not the kind of old box that I feared :-) > Up to now there are usually 2 (max. 3) of the 16 w2k clients switched > on/ connected/ logged in simultaneously. It is not always the same 2 > boxes; this is rather up to the timetable an which class now uses their > 2 or 3 computers in ther class room. > > Because no terminal services are used up to now, the only load -- apart > from the automatic things on tjener -- comes from logging in and off to > samba (from the w2k boxes) and using the squid proxy (very moderately). > > But i think i'll set up a cronjob for checking the load just to be sure. > Thanks for the hint. My guess is that you will find an idle system and the problem gone[1]. I think it is something odd like a power save mode, where disks are spinned down. Everything in cache and only nagios asks ldap to fetch something from disk, which has to spin up. > Looking at the logs so far is even more confusing for me: > Having set up the slapd loglevel to 256 i can see that slapd happily > logs bindings and searches at all times, even the times which are > mentioned by nagios as "CRITICAL". (!?!) AFAIK means CRITICAL from Nagios: got the answer, but not as quick as it should be. > Florian Lohoff said on the german list, that if there is a problem > with bind and it dies it'll write it into daemon.log. But there are > just the normal startup lines from the times i re-/started it by > myself. (I've checked other logfiles as well as daemon.log) Let's call the interesting problem a real challenge > > Confused but nevertheless with best regards, > Patrick Cheers Geert Stappers [1] cronjob prevents disks being spinned down. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
