Hi Olivier,
In fact last week Liverpool had experienced failure of the "BDII"
There it is core time and reability to take reasonable time response
of the globus-mds component, which in fact freze for couple of minutes
and is not relevant to the BDII on whatever machine is running.
in my opinion last upgrade have had lead this problem more visible
with coincidence of just pbs security update without moab/maui at same
time where these services have strong dependencies.
The heavy load has been producing by globus-mds. Some times globus-mds
has exiting leaving infos over log file (see /var/tmp/edginfo-globus-mds.log)
-- log --
Fri Nov 17 21:38:12 GMT 2006 grid-info-soft-register [4222]: log: started daemon PID=4248 "/opt/globus/libexec/slapd"
Fri Nov 17 21:38:12 GMT 2006 grid-info-soft-register [4222]: log: zero registration records
Sat Nov 18 01:05:02 GMT 2006 grid-info-soft-register [4222]: log: daemon PID=4248 terminated, exiting
Cheers
Paul
On Thu, 16 Nov 2006, Olivier van der Aa wrote:
> Dear All,
>
> I would like to know your experience with the BDII stability when there
> is a high load on the CE. What settings of /opt/bdii/etc/bdii.conf are
> you using for BDII_SEARCH_TIMEOUT and BDII_BREATHE_TIME. In some cases I
> have seen that even when the BDII is on another machine it drops the ce
> information published by the mds.
>
> Cheers, Olivier.
>
--
Dr. Paul A. Trepka ;Intl:+44(0)151 794 2137
Oliver Lodge Laboratory ;Fax: +44(0)151 794 3444
Dept. of Physics ;e-mail: [log in to unmask]
The University of Liverpool
Liverpool L69 7ZE
England, UK
|