Hi Dimitris,
Dimitris Zilaskos wrote:
> The slapd daemon in our CE started behaving erratically , as a result
> the site will be down till the issue is resolved.
I guess that you imply the BDII slapd and its surrounding scripts.
In case you run the standard lcg versions of it, they have a bug - or more.
In particular, the default lcg 2.3.0/1 versions have certain race conditions.
Pick the rpm mentioned recently by Laurence Field in the lcg-rollout
and install that one... I believe you won't regret it:
http://lfield.home.cern.ch/lfield/bdii/rpms/lcg-bdii-3.2.0-1.noarch.rpm
We have a version of that latter seriously loaded during the last 24 hours,
and so far it stands strong, reacting reasonably well for a BDII.
We peeked its internals and realised it is *much* better engineered;
what a time waster it wasn't written this way in the first place...
cheers,
Fotis
--
echo "sysadmin know better bash than english" | sed s/min/mins/ \
| sed 's/better bash/bash better/' # Yelling in a CERN forum
|