Tech Meeting Minutes 20101209
Present: Graeme, Sam, Mark, David, Peter, Wahid, Stuart
- WMS - found misconfiguration of BDII on one WMS server (still pointed at svr019). Now fixed. Issues with RAL MyProxy still there, but there seems to be a work around (GGUS 63656).
- WAN rates - going to do testing on Tuesday morning network at risk slot. Use two servers to check they can drive 1Gb.
- glexec - rumbles on. Problems with CMT setup when invoked via glexec.
- svr019 was still appearing in gstat monitoring. Why? Maybe tracked down to bad LGC_GFAL_INFOSYS on WNs. ACTION: David to fix this.
- Change of domain case broke shift.conf. ACTION: Sam to discuss this with J-P B.
- Need to update biomed VOMS certs. ACTION: Peter to fix.
- BDII not yet changed to publish GRIDPP tag. Ongoing.
Actions and Deployment
- glite-MON in svr019 is off.
- Disk deployment -lots of disk deployed. Graeme gave heads-up that ATLAS likely to want to expand DATADISK primarily, retiring other tokens.
- svr026 was throwing memory errors - draining for DIMM replacement
- svr014 performing poorly because of large BLAH registry.
- Could leave us short on the CE front - install svr008 as additional CREAM CE. ACTION on Mark.
- For better ATLAS analysis support need to prevent user analysis jobs from landing on old CV kit. ACTION: Stuart to do this using cluster partitions.
- CREAM server worked! But needs a fix to the 2GB VMEM limit for ATLAS.
- glite-APEL next week.
- Install area nightmare seems finally over.
- Should ask about CVMFS, which ATLAS will push strongly next year.
- Will try to do glite-APEL upgrade early next week.
- Next meeting Dec 16: http://indico.cern.ch/conferenceDisplay.py?confId=117149