Notes from the storage evo meeting 24 Aug 2011 Present: Liverpool: Stephen Lancaster: Matt RAL: Brian, Jens Glasgow: Sam, David QMUL: Chris RHUL: Govind Sheffield: Elena Oxford: Ewan Apologies: Edinburgh: Wahid 0. ATLAS disk usage (Sam, Ewan, all) From ops team meeting yesterday, noted ATLAS lack of use of T2 disk. At Glasgow, ATLAS would have benefited more from Glasgow having bought CPU for some of the money that was used to buy (now unused) storage. Network and bandwidth not being used. Also a poor communications problem? - different things from different people in ATLAS? Disk used to be filled up; also site reliability (as perceived by ATLAS). 1. Canonical install status - should we upgrade pre-1.8.0 DPM sites? - Use of puppet, cfengine, quattor :-) - What about the pre-1.7.0 StoRM sites? Doesn't need to upgrade for 1.8.0 features... no security updates for 1.7.4 and earlier. Most "awesome" reasons for upgrading are in 1.8.2. Should sites wait? Probably best to upgrade to 1.8.0 now, if they can. Can do in a single upgrade, probably: schema changes required. Sam is testing the test release of 1.8.2 at the moment and it seems stable enough: certification is still under way. EMI+UMD - different, because configuration is different. So don't upgrade to this yet. At least do 1.8.X and SL5 first, before doing EMI+UMD. Sites should upgrade to SL5. Note Govind's problems in doing so: MySQL5 migration problems. There are still SL4 security updates. 2. Events and publications I know we discussed it last week, but we didn't come up with anything concrete - there is also HEPiX coming up in Vancouver (24-28 Oct). EGI TF? OGF? 3. Current status of things to document to keep PMB happy: - NFS4 testing/experiences (with dCache, or otherwise) - Plan For Improving (ATLAS) Job Efficiencies (or Monitoring) - Collaborations with EMI 4. AOB ... apologies, I forgot to paste the chat log, since I had to leave early...