Choose timezone
Your profile timezone:
HEPiX IPv6 working group meeting
→
Europe/Zurich
Zoom
Zoom
Description
The timings are approximate.
hepix-ipv6-wg
Zoom Meeting ID
64006864374
Host
Edoardo Martelli
Alternative hosts
Maria Del Carmen Misa Moreira, Costin Grigoras, Andrea Sciabà
Useful links
Join via phone
Zoom URL
HEPiX IPv6 working group - Zoom meeting. 28 July 2022 at 16:00 CEST
Notes by DaveK
Present: Andrea, Bruno, Carmen, DaveK, Duncan, Edoardo, Francesco, Hiro, Martin, Raja
Agreed agenda and issues from last meeting were noted.
1. Work by Carmen and Edoardo to continue (xRoot, FTS - for all/any experiment)
2. Francesco - identify islands of IPv6-only and investigate issues. Make contact with CERN IPv6-only testbed to see what their plans are now.
3. Bruno to update wiki with issues he and KIT have found
4. Continue pressure on dCache updates and dual-stack deployment on WNs
Roundtable updates
CMS (Andrea). configuration of job submission has been modified such that IPv6 is now preferred for HT Condor. Sched-d also moving to dual stack.
KIT (Bruno) - showed slides (attached to agenda). Now seeing increase in IPv6 traffic following dCache upgrades to 7.2.17. Noted that CVMFS_IPFAMILY_PREFER now working too.
Duncan (dCache) - reported he continues to track upgrades to releases that properly prefer IPv6. 39 out of 55 systems are now updated..
Tier 2 status
Andrea reported that the Wuppertal Prod ticket has been closed. All done but still awaiting confirmation by ATLAS.
CMS MIT nodes (xRootD) are now being made IPv6 capable
Study of reasons for use of IPv4 between dual-stack systems (Carmen)
Carmen has no slides today - but work has continued on xrootD transfers. 100% of servers at CERN are dual-stack but some clients are not dual-stack. This is observation of the client and server configurations (and DNS). There is no record of protocol used in the detailed logs of actual transfers. Carmen plans to present more at the next meeting.
Andrea: Does the EOS server have information about protocol used? Andrea will investigate.
Tickets on logging from Duncan (1349, 1693, DMC-1311). New information reported from Mihail. Andrea will ask for more info.
Other issues
We discussed at what point we should ask WLCG MB for a mandate to ask all WLCG Clients to prefer IPv6. Not felt to be appropriate right now as we need to continue work and investigations. We will come back to this.
Francesco reported that he had made contact with Ben Jones at CERN about the IPv6-only testbed. There are no plans to remove the testbed but active work by experiments is still limited, with ALICE for example not yet being ready to test. We wondered what would be required to gain access to WN log files. Hiro reported that they could perhaps set up an IPv6 testbed at BNL.
It was noted that the IPv6 traffic seen on LHCOP/LHCONE at CERN in June had increased significantly, while this same peak is not see in FTS data. Nor is the increase seen in the linecard stats. Is the data flawed in some way? Edoardo will check the July data when that month has ended.
Date of next meetings.
August and September dates/time were changed to maximise attendance of today's participants. There was also concern as to whether the in person meeting planned for Wed 26 October would have to move to 27th Oct (because of earlier request by P4 working group to meet after LHCOPN/LHCONE meeting).
Thursday 25 August meeting was moved to Wednesday 24 August (still at same time 16:00 CEST).
Thursday 29 Sep meeting was moved to Wednesday 28 Sep - and time also changed to 15:00 CEST.
note added on 24 August . P4 working group meeting has been cancelled, so our in-person meeting will be on Wednesday 26 October at CERN from 09:30 to 17:00
There are minutes attached to this event.
Show them.