ROC update
***************
Picked up by COD at yesterday's WLCG/EGEE ops meeting:
Additional information about ticket set to 'Case transfered to political instances':
Site name: UKI-LT2-QMUL
ROC : UKI
Ticket id: 8997 (GGUS id: 40945)
Problem : RGMA-host-cert-valid
'Case transfered to political instances' status from 2008-09-30 and no progress. Now node is in Scheduled Downtime.
Release:
2008-10-18: gLite 3.1 Update 34 was released to production
The update contains:
* LB: A number of non-critical bug fixes (PATCH:1803 , PATCH:2380)
* glite-UI: glite-brokerinfo added
* glite-BDII: obsoleted lcg-info packages removed
fixe for a critical bug which could potentially cause problems on any
node type if the variables BDII_VAR_DIR or BDII_DIR are not set in
the configuration file (not an issue for sites using YAIM). This update also
fixes the problem spotted in the previous release). The release notes given there
are now valid for the glite-BDII service node. In addition the the attributed set_lk_max_locks 2000 has been set in the DB_CONFIG
file which is a partial solution to Bug #42727.
* glite-VOMS (oracle and mysql): obsoleted lcg-info packages removed
* FTS for SL4:
- Some UK sites see ongoing problems with the site report information. Also mentioned by CE ROC.
- IN2P3-CC T1 has now succeeded in configuring its CEs GIP to restrict CMS access to both VOMS:/cms/Role=production and VOMS:/cms/Role=lcgadmin.
- The way the configuration has been made (with help of Steve Traylen) can be found in GGUS ticket #37102.
gLite 3.0 services NOW OBSOLETE
* glite-SE_classic
* glite-VOBOX
* glite-WMS
* glite-PX
* glite-MON
From UK site reports (http://tinyurl.com/6m6bx8):
UCL-HEP:
Many entries in this report do not match the SAM database
UCL-HEP:
Some ops SAM jobs reported failed with user proxy expired. Watching the batch system, I see jobs being ran timely and never waiting in the queues as batch system has had many empty slots throughout. I have also noticed from Steve Lloyd s Atlas tests <http://pprc.qmul.ac.uk/~lloyd/gridpp/atest.html> that jobs are flagged as failed because "Failed to run after 8 Hr 3 Min 12 Sec" which is clearly not the case. I also see that a few other sites have had exactly the same issue as us with Steve s Atlas tests, not only in London.
BHAM-HEP:
- Birmingham had suffered a lot of networking and dns instabilities. Core switch needs rebooting by IS. DNS problem have practically disappeared since IS upgraded BIND from version BIND 9.4.2-P1 to BIND 9.4.2-P2.
- GPFS with software area was made available only midweek by BlueBear/ClusterVision team after a shutdown of over two weeks.
- epgce1 suffered for Marodona errors. I tested scp, globus-url-copy on all the workers, ntp, ... all the tests were successful. I migrated the CE to new server and SL4. Problems seems to have disappeared suddenly. Issue with SL3 CE? Bristol and JET reported lots of Maradona failures too; some were also observed on SL4 CEs -> middleware issue???? Difficult to say at this point, epgce1 hardware was misbehaving at times, will do memtests.
- The scattered software area due to absence of large disk remains an issue. I will relocate it to newly purchase raid.
OX-HEP:
We have had problems with instability of maui.
We have to restart it regularly.
WLCG update
*****************
- GDB rota
Ticket status
***************
https://gus.fzk.de/download/escalationreports/roc/html/20081020_EscalationReport_ROCs.html