Tech Meeting Minutes 20110120
Present: Graeme, Sam, David, Wahid, Stuart, Andy
- Durham have a continuing problem with biomed VOMS setup. Will rerun YAIM to see if this fixes the issue.
- ECDF CREAM CE working well - hooks which were needed will go into the next CREAM release.
- SAM tests are failing because they go to the wrong queue(?)
- 6GB VMEM now requested.
- ACTION Graeme to send pilots.
- ECDF glite APEL done.
- ECDF not running LHCb pilots.
- Graeme found that the VO view was showing 444444 waiting jobs, so there is a publishing problem to be fixed. ACTION Andy/Wahid.
- ECDF cvmfs
- Have spoken to Orlando. He like the idea of cvmfs, but control would be on the systems team. This might cause operational problems. Alternative strategy is to mount cvmfs on current s/w server and export this via nfs.
- ECDF squid
- A VM is being commissioned for this.
- Durham - Graeme noted that they are now in the GRIDPP project in gstat.
- Glasgow - Downtime next week for card replacement in the new disk servers. ACTION Graeme to get in touch with Cedric about declaring the unavailable files to be 'suspicious'.
- Glasgow disk - Sam is in the process of migrating to SL5 (most of the Viglen servers are still SL4). fs choice is ext4 with some tuning applied.
- Security Incident Handling
- People have looked at the document. Communication is clearly very important. Noted the need for flexibility in local handling, e.g., at ECDF we might have to shut a box down if physical network disconnection is impractical.
- Next meeting: Jan 27, http://indico.cern.ch/conferenceDisplay.py?confId=124890