Hi all,
continuing with our new CE configuration, we've found a new strange
behaviour: Sometimes cdf vo appears as "supported":
[arnaubria@ui01 ~]$ ldapsearch -x -H ldap://ce05.pic.es:2170 -b mds-vo-name=resource,o=grid|grep -i cdf
# cdf, ce05.pic.es:2119/jobmanager-lcgpbs-gshort, resource, grid
dn: GlueVOViewLocalID=cdf,GlueCEUniqueID=ce05.pic.es:2119/jobmanager-lcgpbs-gs
GlueVOViewLocalID: cdf
GlueCEAccessControlBaseRule: VO:cdf
GlueCEAccessControlBaseRule: VO:cdf
GlueCEAccessControlBaseRule: VOMS:/cdf/Role=software
GlueCEAccessControlBaseRule: VOMS:/cdf/Role=production
# cdf, ce05.pic.es:2119/jobmanager-lcgpbs-glong, resource, grid
dn: GlueVOViewLocalID=cdf,GlueCEUniqueID=ce05.pic.es:2119/jobmanager-lcgpbs-gl
GlueVOViewLocalID: cdf
GlueCEAccessControlBaseRule: VO:cdf
GlueCEAccessControlBaseRule: VO:cdf
GlueCEAccessControlBaseRule: VOMS:/cdf/Role=software
GlueCEAccessControlBaseRule: VOMS:/cdf/Role=production
But sometimes not:
[arnaubria@ui01 ~]$ ldapsearch -x -H ldap://ce05.pic.es:2170 -b mds-vo-name=resource,o=grid|grep -i cdf
[arnaubria@ui01 ~]$
We do NOT support cdf:
# ls
groups.conf nodes site-info.def users.conf vo.d wn-list.conf
# grep -r cdf *|grep -v users.conf
It's not jsut a cdf problem, it should be a gip missconfiguration cause
we also find the problem with totalcpu:
[arnaubria@ui01 ~]$ ldapsearch -x -H ldap://ce05.pic.es:2170 -b mds-vo-name=resource,o=grid|grep -i totalcpus
GlueCEInfoTotalCPUs: 780
GlueCEInfoTotalCPUs: 780
GlueCEInfoTotalCPUs: 780
GlueCEInfoTotalCPUs: 780
[arnaubria@ui01 ~]$ ldapsearch -x -H ldap://ce05.pic.es:2170 -b mds-vo-name=resource,o=grid|grep -i totalcpus
GlueCEInfoTotalCPUs: 904
GlueCEInfoTotalCPUs: 904
GlueCEInfoTotalCPUs: 904
GlueCEInfoTotalCPUs: 904
(it must be 780, we have modified /opt/lcg/libexec/lcg-info-dynamic-pbs
for taking only pic nodes from a shared pbs).
So, where are we making the mistake? What should we start looking at?
TIA,
Arnau
|