On 23/03/2010 17:04, Wahid Bhimji wrote:
> Not all the numbers are 0 though,
> GlueSALocalID: atlasscratchdisk:replica:online
> GlueSATotalOnlineSize: 318805
> GlueChunkKey: GlueSEUniqueID=se01.esc.qmul.ac.uk
>
> and some of these look more like they "should be" (I think - need to dig the document out)
>
> PS its ldapsearch -x -H ldap://se01.esc.qmul.ac.uk:2170 -b o=grid
Here's what I use:
ldapsearch -x -LLL -H ldap://${LCG_GFAL_INFOSYS}/ -b
glueseuniqueid=se01.esc.qmul.ac.uk,mds-vo-name=uki-lt2-qmul,mds-vo-name=local,o=grid
... which works in case you cannot access the GRIS (some people would
firewall them off and just expose the site BDII.)
The unit for things like GlueSATotalOnlineSize is GB, so the number is
319 management terabytes. The legacy ones are in KB and also look right.
Running the flavian tests, I notice the following errors in the SAs:
Error: GlueSAReservedOnlineSize set and correspondent GlueVOInfo object
does not have a GlueVOInfoTag entry.
and warnings:
Warning: GlueVOInfo object for VO dteam defined for
GlueSALocalID=dteam:replica:online
Warning: GlueSAPath is not mandatory
Warning: This host does not run an SRM v1.1 service
...which to be honest I would not worry about, I think the CASTOR
information provider pretty much raises the same errors and warnings.
More curiously, the VOInfo test reports
Error: No GlueVOInfo objects defined for this SE
which is odd because I can definitely see them, and the SA test picked
them up.
Overall, I think it looks really good. I haven't seen what it looks
like with the new tests but I think everything looks universally
terrible anyway.
Cheers
--jens
|