Re: LDAP server scaling problem?
2011/11/22 Petter Reinholdtsen <pere@hungry.com>:
>
> [Petter Reinholdtsen]
>> If I got the details right, a Linux process can by default only have
>> 1024 files open. This can be adjusted using 'ulimit -n', but I am not
>> sure if it will work with select() calls because of a hardcoded
>> constant in the system header files.
>
> I've been told that it work just fine to extend it for slapd. It should
> thus work to add 'ulimit -n 8192' in /etc/default/slapd.
>
>> Anyone else seen similar problems? Is there any reason not to include
>> "idletimeout 60" or some other sensible timeout in our default
>> slapd.conf?
>
> No-one else experience problems like this with Debian Edu/Lenny?
>
> I see a problem with OpenLDAP spending 100% CPU and not responding to
> requests for a long time. My local OpenLDAP expert sent me to
> <URL:http://www.openldap.org/lists/openldap-software/200212/msg00435.html>
> and <URL:http://www.openldap.org/its/?findid=6856> which might be
> relevant. I have not had time to test yet. Anyone else experience
> hanging OpenLDAP in Lenny?
>
As you said in a previous email, in Extremadura we had the same
problems. With the iddletimeout option they were fixed.
You need to have many concurrent users to notice the problems ( aprox.
more than 100), and I guess that's why there are not more people
complaining, as this concurrency is not very usual.
Since May we have updated our ldap servers to Squeeze, and we have had
no problems up to now, with idletimeout=0 .
Regards.
Reply to: