Andrea Chierici wrote:
> Maarten Litmaath wrote:
>
>> Ricardo Graciani wrote:
>>
>>> Hi Andrea,
>>>
>>> you are publishing:
>>>
>>> GlueHostBenchmarkSF00: 400 <===========
>>> GlueHostBenchmarkSI00: 380 <===========
>>> GlueHostMainMemoryRAMSize: 2015
>>> GlueHostMainMemoryVirtualSize: 2532
>>> GlueHostNetworkAdapterInboundIP: FALSE
>>> GlueHostNetworkAdapterOutboundIP: TRUE
>>> GlueHostOperatingSystemName: Redhat
>>> GlueHostOperatingSystemRelease: 2.4.20-30.7.legacysmp
>>> GlueHostOperatingSystemVersion: 1 SMP Fri Feb 20 10:12:55 PST 2004
>>> GlueHostProcessorClockSpeed: 2395
>>> GlueHostProcessorModel: Intel(R) Xeon(TM) CPU 2.40GHz
>>>
>>>
>>> We (LHCB) are currently using this value to normalize the CPU
>>> time
>>> request to all sites.
>>>
> Sorry, I can't understand one thing... is this a problem??
Ricardo has answered that. Please fix your /opt/edg/var/etc/ce-static.ldif;
note that that file is generated by the "ceinfo" object. In your site-cfg.h
the relevant variables are CE_IP_SI00 and CE_IP_SF00. Note that CE_IP_SF00
should be set to zero, as recently discussed on the rollout list.
>>> Furthermore your queue times seem to be still in seconds, after
>>> the upgrade they should be in minutes (there is an LCG object that takes
>>
>>
>>
>> /opt/edg/libexec/edg-ce-all from RPM edg-info-ce-lcg2.6.38-1.
>>
> We are using a "patched" version of edg-ce-all to fix the counting of
> running job on each queue we publish. This was done with the help of
> David Smith, and we did not upgrade it to the latest release, sorry. Is
> it enough to fix the queue time or does the new file include other
> differences?
The new RPM should already include your changes. Save the old edg-ce-all,
install the new RPM and do a diff.
>> > Don't forget after upgrading the CE to make sure that the experiment
>> > specific runtime environment tags can still be published. For this move
>> > the /opt/edg/var/info/<VO-NAME>/<VO-NAME>.ldif files to <VO-NAME>.list.
>>
>> The files used to be called "$VO.ldif", but since they do not contain
>> data in LDIF format, the names were changed to "$VO.list".
>> For example:
>>
>> /opt/edg/var/info/lhcb/lhcb.list
>>
>> On wn-04-07-02-a.cr.cnaf.infn.it even the old files are gone.
>>
> After reinstalling our ce from scratch, we did not find ANY file inside
> those dirs. Where can I get them? This is quite strange...
If you installed from scratch (!), it is no surprise they are gone!
It seems Ricardo and other VO "sgm" users will have to regenerate them.
|