You were right ;-)
Thank You very much. Indeed I forgot that this is user edginfo who runs
all the scripts on behalf of slapd.
Here's what I get when I run lcg-info-wrapper as edginfo:
[edginfo@ce var]$ /opt/lcg/libexec/lcg-info-wrapper |grep CPU
sh: line 1: /opt/lcg/var/gip/tmp/lcg-info-dynamic-pbs.ldif.28820:
Permission denied
sh: line 1: /opt/lcg/var/gip/tmp/lcg-info-dynamic-software.ldif.28822:
Permission denied
mv: cannot stat `/opt/lcg/var/gip/tmp/lcg-info-dynamic-pbs.ldif.28820': No
such file or directory
mv: cannot stat
`/opt/lcg/var/gip/tmp/lcg-info-dynamic-software.ldif.28822': No such file
or directory
GlueCEInfoTotalCPUs: 0
GlueCEStateFreeCPUs: 0
GlueCEPolicyMaxCPUTime: 0
GlueCEInfoTotalCPUs: 0
GlueCEStateFreeCPUs: 0
GlueCEPolicyMaxCPUTime: 0
GlueCEInfoTotalCPUs: 0
GlueCEStateFreeCPUs: 0
GlueCEPolicyMaxCPUTime: 0
--
Piotr Siwczak <[log in to unmask]>
System Administrator
Poznan Supercomputing and Networking Center
Supercomputing Department
On Thu, 17 Mar 2005 [log in to unmask] wrote:
> On Thu, 17 Mar 2005, Piotr Siwczak wrote:
>
>> Hi,
>>
>> I investigated the problem in more detail and I found some more
>> information is reported incorrectly in my case. It turned out I miss some
>> info about the number of available resources too.
>>
>> Some facts about my site GIIS at first:
>>
>> -> we are using lcg-info-wrapper as the provider script
>> -> we are running Itanium LCG version from Openlab
>> -> we are using GIIS (port 2135)as the standard info provider (we don't
>> have site BDII)
>>
>> Right after the start of globus-mds, ldapsearch returns correct results.
>>
>> Here's what I get about CPUs when running ldapsearch some time later :
>>
>> ldapsearch -x -H ldap://ce.egee.man.poznan.pl:2135 -b
>> Mds-Vo-name=egee.man.poznan.pl,o=grid |grep CPU
>> GlueCEInfoTotalCPUs: 0
>> GlueCEStateFreeCPUs: 0
>> GlueCEPolicyMaxCPUTime: 0
>> GlueCEInfoTotalCPUs: 0
>> GlueCEStateFreeCPUs: 0
>> GlueCEPolicyMaxCPUTime: 0
>> GlueCEInfoTotalCPUs: 0
>> GlueCEStateFreeCPUs: 0
>> GlueCEPolicyMaxCPUTime: 0
>>
>> But when I run lcg-info-wrapper by hand:
>
> Try running it as user "edginfo". You may have a problem with permissions
> on files or directories.
>
>> [root@ce root]# /opt/lcg/libexec/lcg-info-wrapper |grep CPU
>> GlueCEInfoTotalCPUs: 4
>> GlueCEStateFreeCPUs: 4
>> GlueCEPolicyMaxCPUTime: 720
>> GlueCEInfoTotalCPUs: 4
>> GlueCEStateFreeCPUs: 4
>> GlueCEPolicyMaxCPUTime: 15
>> GlueCEInfoTotalCPUs: 4
>> GlueCEStateFreeCPUs: 4
>> GlueCEPolicyMaxCPUTime: 2880
>>
>> This indicates that there's an issue with slapd daemon itself, and not
>> with the infoprovider scripts.
>>
>>
>>
>>
>>
>> --
>> Piotr Siwczak <[log in to unmask]>
>> System Administrator
>>
>> Poznan Supercomputing and Networking Center
>> Supercomputing Department
>>
>> On Tue, 15 Mar 2005, Laurence wrote:
>>
>>> The error is that the GlueClusterUniqueId in the two different dns do
>>> not match.
>>>
>>> dn: GlueClusterUniqueID=ce.egee.man.poznan.pl,
>>> Mds-Vo-name=egee.man.poznan.pl,o=grid
>>> dn: GlueSubClusterUniqueID=ce.egee.man.poznan.pl,
>>> GlueClusterUniqueID=egee.man.poznan.pl,
>>> Mds-Vo-name=egee.man.poznan.pl,o=grid
>>>
>>> In the GlueCluster entry
>>> GlueClusterUniqueID=ce.egee.man.poznan.pl
>>> In the GlueSubCluster entry
>>> GlueClusterUniqueID=egee.man.poznan.pl
>>>
>>>
>>> Laurence
>>>
>>>
>>> Jeff Templon wrote:
>>>
>>>> Piotr Siwczak wrote:
>>>>
>>>>> GlueClusterName: ce.egee.man.poznan.pl
>>>>> GlueClusterUniqueID: ce.egee.man.poznan.pl
>>>>> dn: GlueSubClusterUniqueID=ce.egee.man.poznan.pl,
>>>>> GlueClusterUniqueID=ce.egee.
>>>>
>>>> looks like the above is missing the poznan.pl piece.
>>>>
>>>>> GlueChunkKey: GlueClusterUniqueID=ce.egee.man.poznan.pl
>>>>> GlueSubClusterName: ce.egee.man.poznan.pl
>>>>> GlueSubClusterUniqueID: ce.egee.man.poznan.pl
>>>>
>>>
>>
>
|