PIC Tier-1 Site Report (12/3/07)
================================

We are experiencing network problems in the CERN-PIC connection from thursday 8th March in the morning. The symptom is that a high fraction of the connections timeout. We contacted first with RedIris (spanish NREN) but they do not see anything strange on their side and the problem is still there today (12th March). We have opened GGUS-19568 to see wether the Network operations team at CERN could help us to debug the problem.

-lcg-CE service:
 +We closed the monitoring flag for our CE ce04.pic.es last week but we continued to receive ggus tickets concerning job submissions failed in this machine. It seems we have to wait more than 3 days and not to publish the information about this CE in order to not to be monitored by SAM. 

-SRM-tape service: We changed the srm version we were publishing from 1.1.1) to 1.1.0 because we received alarms in the GSTAT tests.

-SRM-disk service: The certificate of one of the dCache pool nodes expired on the 2nd March. This was one of the 1st dCache nodes deployed, and had a problem in the sensor monitoring the certificate lifetime. This problem caused the SRM-disk SAM tests to fail for the whole weekend. We tried to remove the machine from the dCache pool on friday, but a problem in the automatic recovery actions system, put it back to production. The new certificate was finally received and the problem solved on monday the 5th March.

-glite-CE service:
 +We still get a lot of errors in our gLiteCE with the error: "The job attribute PeriodicHold expression ''Matched =!= TRUE && CurrentTime > QDate + 900'' evaluated to TRUE". From our point of view jobs finish properly, we are not able to reproduce the problem. 
 +We had a punctual problem with our gLiteCE. The load reached almost 200 because  of the number of glite-gatekeepers that were running. After a reboot the problem was solved.