INFORMATION SERVICES DIVISION Computational Resource Allocation Group (CRAG) Monthly Meeting Friday 20th November 2015, 13:00 Room 103, Podium Building, 1 Eversholt Street, London NW1 2DN Chair: Dario Alfe [DA] Present: Sergey Yurchenko (SY]; Andrew Martin (AM); Javier Herrero (JH); Owain Kenway (OK) Apologies: Thomas Jones (TJ); Clare Gryce (CG); Michail Stamatakis (MS); Nikolas Kaltsoyannis (NK); Ian Kirker (IK); Jo Lampard (JL); Tom Couch (TC); Nick Achilleos (NA); William Hay (WH); Ton Jones (TJ); Michail Stamatakis (MS) In attendance: Kerry Deane (KD) - Minutes MINUTES 1. Approval of Minutes of last meeting held on Friday 18th September 2015 [Chair] The Group approved the Minutes of the last meeting. There were no matters arising. 2. Update on status of current Actions (Chair/All) The list of current Actions (see table at end) was updated. 3. Review of any requests for additional resources on local HPC facilities (Chair/All) Ana Ferreira requested an extension of her previous request of 7th October 2014 for an increased Scratch quota of 2 TB and this was approved. Federico Minneci requested his Scratch space be increased to 5 TB for work on bioinformatics projects. Patrick Lombard requested her Scratch space be increased to 5 TB for work on projects with sequencing data. Page 1 of 6 4. Review of Legion usage statistics (OK) Everything except GPU nodes were moved up until this week. Due to a bug that was discovered this week, the stats are for the old part of the cluster New users are experiencing longer wait times but hopefully people can be moved to Grace. The longest that a newly submitted job will wait is 12 hours. There are now 6 type T nodes. The advantage of Grace is that it is based within UCL and it has no legal limits ACTION: OK to investigate what happens if somebody buys nodes and subsequently leave UCL. 5. Review of EMERALD usage statistics (OK) Emerald is up and running. 6. Update report on Cfl and new HPC (OK) Grace has passed all acceptance tests and we are addressing a minor issue. People have access to their data on Legion, which was not originally intended. There was a CI with DFS and authentication service was impacted, but this is being addressed. Applications such as LAMPS will be rebuilt and then access to them will be given to users. If there are no further issues, they will be available at the end of this month. We may have to look at storage infrastructure for Legion next year. ACTION: OK to give people a script so that they can move data across to the new system. 7. AOB Other names apart from Grace and Legion are needed. The Group can think about this for the next meeting. It would also be useful to have names for the machines. ACTION: All to think about alternative names for Grace and Legion for discussion at the next meeting. ACTION: OK to circulate text for citing Legion to this group. 8. Next meeting date Friday 18th December 2015 from 13.00-15.00. Venue: Room 103, 1st floor, Podium Building, 1 Eversholt Street, London, NW1 2DN. 9. Agenda (Items) for the next meeting: Standing items: 1. 2. 3. 4. 5. Approval of Minutes of last meeting. Update on status of current Actions. Review of any requests for additional resources on local HPC facilities. Review of Legion usage statistics. Review of EMERALD usage statistics. Page 2 of 6 6. Update report on CfI and new HPC (from CG). 7. Data repatriation plans. CURRENTLY APPROVED EXCEPTIONAL REQUESTS Requesting CRAG approval user date details of exception start date end date agreed agreed Implementation date removed Notes (for info) Conde 17/07/2015 Scratch quota increased to 3TB Pikoula 17/07/2015 8 U-type nodes 17/07/2015 13/08/2015 for a month 18.09.2015 Agreed to extend for another month Meng 09/09/2014 Scratch quota increased to 3TB and extended 09/09/2014 30/09/2015 18.09.2015 Agreed to extend for another year Herrero 09/09/2014 Scratch quota increased to 6TB and extended 09/09/2014 30/09/2015 19.09.2015 Agreed to extend for another year Ferreira 07/10/2014 Scratch quota increased to 2TB and extended 07/10/2014 31/10/2015 18.09.2015 OK to contact user closer to the expiry date. 17/07/2015 16/07/2016 20.11.15 Agreed to extend for another year. Wang 20/02/2015 Scratch quota increased to 2TB and extended 08/01/2015 31/01/2016 Wellington 23/02/2015 Scratch quota increased to 1TB for the period 23/02/2015 01/08/2015 Bloch 20/03/2015 Scratch quota increased to 1TB and extended 20/03/2015 31/03/2016 Page 3 of 6 Advise user to reapply upon expiry as had requested longer timeframe Requesting CRAG approval user date Oostra details of exception 22/05/2015 Scratch quota increased to 3TB and extended start date end date agreed agreed Implementation date removed Notes (for info) 22/05/2015 31/05/2016 22/05/2015 40+ nodes to 16/06/2015 23/06/2015 Terras/ be reserved for Hetherington JISC Data Spring project events Javed 23/10/15 Additional disk 10/2015 space required for MRC DTP: structural study of the role of the ribosome in co-translational protein folding during protein biosynthesis. 06/2017 Minneci 20/11/2015 Personal 23/11/15 scratch space increased to 5 TB. Working on bioinformatics projects. 22/11/16 Lombard 20/11/15 Scratch space 10/11/15 increased to 5 TB. Working on projects with sequencing data. 09/11/16 Advise user to reapply upon expiry as had requested longer timeframe Between the hours of 09:0018:00 on both dates only LIST OF CURRENT ACTIONS Shaded (closed/completed) items will be deleted in the next version. 166 Actions Status Owner Re-application process, stats for re-application and project account/ publication data (09/12/14) NEW ACTION: Check with RAGG to add to backlog and confirm usage. Report back to CRAG with plan on how to implement. OK (20/02/15) ONGOING: Contractor has been hired. Page 4 of 6 G/OK Actions Status Owner (20/03/15) ONGOING: OK to check on progress. (22/05/15) ONGOING (19/06/15) ONGOING (17/07/15) ONGOING: New web form is almost complete and will be approved internally before it is presented at the next CRAG meeting. (18/09/15) ONGOING: OK and CG have reviewed the data with the developer and confirmed that it is almost complete. Slight changes to be made to how the funding information is gathered. Approval for renewal granted. (23/10/15) ONGOING 169 Legion usage statistics (20/11/15) ONGOING: As Helen is leaving, JH is currently recruiting a software developer who has the necessary skills for this. (17/04/15) NEW ACTION: OK to talk to OK Systems team re metrics/file system performance. (22/05/15) ONGOING (19/06/15) ONGOING (17/07/15) ONGOING (18/09/15) ONGOING. Currently waiting for what packages come with Grace in terms of statistics gathering. (23/10/15) ONGOING: OK is putting fixes in place. Legion and Grace will be separate but the same system. (20/11/15) ONGOING: Legion and Grace are still separate but will hopefully be on the same system by the next meeting. 173 CfI (19/06/15) NEW ACTION: CG to alert EMERALD user community to the possibility of extending the service into 2016/17 if users can find funding. (17/07/15) ONGOING (18/09/15) ONGOING: Have not received any updates from CG. Page 5 of 6 CG Actions Status Owner (23/10/15) ONGOING 20/11/15) ONGOING: No updates. 176 STFC Statistics reporting quality (17/07/15) NEW ACTION: OK to contact STFC to discuss/ identify a better method of obtaining stats reports. OK (18/09/15) ONGOING: CFI have stopped providing statistics; OK has been attempting to contact them via email. (23/10/15) ONGOING: OK will continue to investigate. (20/11/15) ONGOING: No statistics this month as the server was down. 177 New Grace Scheduling policy (17/07/15) NEW ACTION: OK to ensure new policy is implemented in Grace prior to rollout. OK (18/09/15) ONGOING (23/10/15) ONGOING: Interactive access with a time limit was previously agreed. 178 Legion user statistics 179 Making Iridis home directories available on Grace 180 Decide on text for citing the use of RC Platforms Services 181 Circulate current wording for 180 (20/11/15) ONGOING: William has made implementations already but needs to test it. (20/11/15) NEW ACTION: OK to investigate what happens if somebody buys nodes and subsequently leave UCL. (20/11/15) NEW ACTION: OK to give people a script so that they can move data across to the new system. (20/11/15) NEW ACTION: All to think about the language used for citing the use of our services (due to the fact that there will now be both Grace and Legion) for discussion at the next meeting. (20/11/15) NEW ACTION: OK to circulate current text for citing Legion to this group. Page 6 of 6 OK OK All OK