Hi Maarten,
On 07/31/2010 05:34 PM, Maarten Litmaath wrote:
> Hallo Christoph,
>
>
>> we run into the same problem with upgraded WMS again. Now I believe I
>> understood the problem. In /var/local/condor/log/GridmanagerLog.glite
>> many restarts of the component were reported after such a crash:
>>
>> 07/31 16:40:04 [18921] gahp server not up yet, delaying ping
>> 07/31 16:40:04 [18921] GAHP server pid = 19331
>> 07/31 16:40:04 [18921] gahp->nordugrid_ldap_query returned -101 for
>> resource korundi.grid.helsinki.fi
>> 07/31 16:40:04 [18921] ERROR "nordugrid_ldap_query failed!" at line 211
>> in file nordugridresource.cpp
>>
>> Searching a bit the internet I found that you got trapped with a similar
>> problem in May:
>>
>> http://lindir.ics.muni.cz/pipermail/egee-jra1/2010-May/012580.html
>>
>> When I remove the old nordugrid_gahp and use the one included in
>> Condor-7.4 things start to work again.
>>
> YAIM should have configured the new nordugrid_gahp automatically.
>
>
For us it did not:
> grep nordugrid /opt/glite/yaim/functions/config_condor_wms
setValue NORDUGRID_GAHP "/opt/glite/sbin/nordugrid_gahp"
> rpm -qf /opt/glite/yaim/functions/config_condor_wms
glite-yaim-wms-4.0.7-1.noarch
Should be the correct version.
> The old glite-condor-extra rpm has been removed from the WMS metapackage.
>
>
This I can confirm, although the RPM remind on the system after yum update.
> I presume the nodes have been reconfigured?
> Were there any suspicious messages in /opt/glite/yaim/log/yaimlog?
> Do you have any local customizations, e.g. local YAIM functions?
>
Nothing in the logs and no local function in that area.
> If not, please open a WMS bug.
>
>
I will open a GGUS ticket.
> At CERN we do have a local customization in this area, so we did not
> verify if the standard YAIM does the correct thing.
>
For me it appears to be buggy.
Cheers, Christoph
|