|Main||EGI.eu operations services||Support||Documentation||Tools||Activities||Performance||Technology||Catch-all Services||Resource Allocation||Security|
Detailed agenda: Grid Operations Meeting 10 October 2011 14h00 Amsterdam time
|EVO direct link|| Pwd: gridops|
|EVO details||Indico page|
1. Middleware releases and staged rollout
1.1 EMI-1 release status
External twiki page with status update: EMI update for the EGI Grid Operations Meeting
1.2 WMS 3.3.4
A new WMS release (WMS 3.3.4) is being prepared. It is tracked in this savannah bug: https://savannah.cern.ch/task/?22847
where the list of addressed RFCs is listed:
https://savannah.cern.ch/bugs/?86485 https://savannah.cern.ch/bugs/?86490 https://savannah.cern.ch/bugs/?81376 https://savannah.cern.ch/bugs/?81568 https://savannah.cern.ch/bugs/?83062 https://savannah.cern.ch/bugs/?85327 https://savannah.cern.ch/bugs/?86682 https://savannah.cern.ch/bugs/?86769 https://savannah.cern.ch/bugs/?86956 https://savannah.cern.ch/bugs/?87259
So this new WMS addresses also bug #86490 (EMI WMS stops accepting jobs after 31999th submission for the same DN (on ext3)) which was the main reason for not accepting the WMS in UMD. However there is another problem, tracked in this [https://ggus.eu/tech/ticket_show.php?ticket=75083 ggus ticket].
Basically gridsite has changed the structure of the proxy cache in EMI. In glite there was a flat structure; all proxy files get stored in: /var/glite/proxycache
In EMI, instead, there is a different structure; a proxy file gets created in:
Because of WMS Savanah bug #86490, expired proxy files (and the related directories) didn't get cleaned at all in the WMS. This means that on ext3 (where there can be at most 31999 sub-directories) the WMS couldn't accept anymore jobs for a certain user after 31999th submission for that user.
Even with the fix of that bug there are at any rate problems on ext3:
a: there can be at most 31999 different users submitting to a certain WMS b: there can be more or less 32000 valid (i.e. not expired) proxies for a certain user (I say more or less because the purging of the expired proxy files is done on the WMS by a cron job which runs every 6 hours)
A ticket for this problem (in particular the second one can be an issue) has been filled (it is assigned to the gridsite SU since the issue should be addressed in gridsite).
In the meantime the workaround that can be used to address this issue is to use ext4 when installing a WMS node (on ext4 the limit is increased to 64000 subdirectories).
1.3 Staged Rollout (Mario)
WIKI of staged rollout procedures has been updated to include:
- The status of EMI components can be found here:
1.3.1 gLite 3.2
- In staged rollout
- DPM 1.8.2: not yet accepted by any EA - There are still many sites with DPM in glite 3.2
- LFC 1.8.2: not yet accepted by any EA - There are still many sites with DPM in glite 3.2
Action on MDavid to send a mail to interested parties, EAs and SA1.
Components from EMI:
- WMS 3.3.3: new version upcoming, discussed earlier
- Site BDII 1.0.1: several EAs
- Top BDII 1.0.1: 1 EA but expecting at least one more
- ARGUS 1.4.0: several EAs
- gLexec 1.0.1: 1 EA but expecting at least one more
Components from IGE (Globus) all with 1 EA doing the test:
- RLS 5.0.4
- GridFTP 5.0.4
Components in verification status or to be verified:
- AMGA 2.1.2: has already a committed EA for staged rollout
- FTS 2.2.6: pending commitment of EAs
- UNICORE 6.4.1:
1.3.3 Candidate list for UMD1.3
UMD 1.3 is scheduled for the 31st October. The candidate products are:
- TopBDII 1.0.1
- SiteBDII 1.0.1
- ARGUS 1.4.0
- gLexec 1.0.1
- UNICORE 6.4.1:
- Globus 5.0.4
1.3.4 Candidate list for UMD1.4
The UMD 1.4 is scheduled for the end of January 2012, as such if a decision is reached by OMB for early inclusion of WMS 3.3.4 into UMD, then a re-scheduling is necessary. Storm 1.7.2 may profit from this update and be included as well.
2. Operational Issues
gLite 3.2 critical bugs and patches
New How-To to be reviewed
A new how-to is available in the documentation section of the wiki page: Change the Site-BDII how-to
It is a very simple step-by-step procedure, please review it, and provide comments or suggestions (if any).
Thanks to A.Paolini for sharing this manual.
3.3 Next meetings
24 October 2011