Re: slapd stuck in sched_yield() call

["Dieter Kluenter" <dieter@dkluenter.de>]

Misty Stanley-Jones <misty@borkholder.com> writes:

> Hi all,
>
> I've got an openLDAP read-only replica running on a server, and I noticed 
> today that it is using 90-95% of the CPU and it is looping in sched_yield() 
> system call (according the strace).  It really should not be doing anything 
> as no changes are being made to the master LDAP repository at all.  Could 
> this be a bug?  I am running 2.2.6 (the latest RPM from SUSE) on this system.  
> As it is only a failover system I'd rather have an easy-to-manage RPM than 
> the latest and greatest.  But maybe this is a known bug that I am running 
> into?  It would only concern me because it seems to be pegging the CPU.

You should update your SuSE distribution, 9.0 is quite old :-)
There are a few questions to answer prior to a cure

What is your database definition? (ldbm, bdb)
What is the size of your database? (number of entries or in MB)
What is the contents of your DB_CONFIG?
Has your replica being down?
Did you try a db_recover?

-Dieter

-- 
Dieter Klünter | Systemberatung
http://www.dkluenter.de
GPG Key ID:01443B53