- Compact style
- Indico style
- Indico style - inline minutes
- Indico style - numbered
- Indico style - numbered + minutes
- Indico Weeks View
https://tinyurl.com/T1-GGUS-Open
https://tinyurl.com/T1-GGUS-Closed
https://lcgwww.gridpp.rl.ac.uk/utils/availchart/
https://cms-site-readiness.web.cern.ch/cms-site-readiness/SiteReadiness/HTML/SiteReadinessReport.html#T1_UK_RAL
http://hammercloud.cern.ch/hc/app/atlas/siteoverview/?site=RAL-LCG2&startTime=2020-01-29&endTime=2020-02-06&templateType=isGolden
* RT: #296706: Optimise XRootD checksums for TPC transfers from ECHO
* Closed long-standing GGUS:145510 Stage-in/stage-out:
Several contributing factors in improvements; change to pilot code appears to have reduced finally the job error rate.
Still observe significance differences in errors rates between SSD and non-SSD (for the staging failures), but all at <1%
* Sub-quota values;
- Aim to remove additional layers of prioritisation with change to Atlas sub-group quotas. Did not work, other VO's manage to absorb slots in negotiation cycles?
- ATLAS currently 65% (nominal 109%)
- CMS 250%. (nominal 160% )
- LHCb 153% (nominal 140%)
- ALICE 194% (nominal 4 – 140%)
* ATLAS may exceed (overall) 2020 tape pledge
* Will plan to clean up and delete secondary tape replicas (when is convenient for RAL)?
* Also discussions ongoing on file size.
* ceph - xrootd plugin fchmod fix (GLA)
* TPC Status
- XrootD; Smoke tests passing for RAL-LCG2(prod) and RAL-CEPH(test).
stress test
- checksum 'fix' needed: (RT: #296706)
- HTTP: Using the test gateway;
- Main problems around authorisation with gridmap-file
- setup with "http.gridmap", works with Firefox download, but not curl or davix-get
- (VOMS would use http.secxtractor libXrdHttpVOMS.so)
- Trying a patched version of libXrdHttpVOMS.so (to use the grid map file), but no success yet
-- Other possibility to enable VOMS for testing of http?
- Is this complicated
No particular problems with operations this week. Job efficiency has been a bit up and down, but I have not found a particular cause yet.
The CMS 'batch allotment' is not accepting jobs; Jose has been looking into it.
I have also been studying the SSD vs non-SSD jobs. SSDs appear to cause fewer failures due to error code associated with FileRead. Both WN types have similar failure rates with FileOpen type errors.
LHCb:
DUNE: