Choose timezone
Your profile timezone:
HEPiX IPv6 working group meeting
→
Europe/Zurich
Vidyo
Vidyo
HEPiX IPv6 working group meeting - Vidyo - Thursday 25 February 2016
Present: Marian Babik, Jiri Chudoba, Alastair Dewhurst, Thomas Finery, Terry Froy, Kashif Hafeez, Bruno Hoeft, Tiju Idiculla, David Kelsey, Fernando López Muñoz, Edoardo Martelli, Raja Nandakumar, Francesco Prelz, Duncan Rand, Andrea Sciaba, Dan Traynor, Ulf Tigerstedt.
Apologies: Kars Ohrenberg, David Mitchell
Introduction
The agenda was agreed.
Minutes of 21-22 January 2016 F2F meeting
No corrections needed. Accepted.
Review of actions and ongoing issues
1. AlastairD. Work on IPv6-only document together with other experiment reps.
Ongoing. Alastair has emailed the other reps. There has been some discussion but work continues.
2. AlastairD. Get one slide input to the ATLAS site jamboree and ask for longer session at ATLAS s/w week.
Done. Alastair has a plenary talk next week and has asked for a parallel session in the June ATLAS Computing and software week.
3. AlastairD. Work on testing IPv6-only WN.
Ongoing.
4. DaveK. See if HEPiX is interested in IPv6 training in April.
Done. It did not fit this time because of too many other additional sessions.
5. DaveK. Distribute draft IPv6 security guidance for more feedback.
Ongoing.
6. BrunoH. Update LHCONE IPv6 wiki.
Ongoing. Will do this during the next 2 weeks.
7. UlfT. Take on maintenance of checking IGTF CA CRL status re IPv6.
Ongoing. Work in progress.
8. AndreaS. EOS - can they deploy dual stack on production services?
Done. Asked. This is not possible right now. They plan to do this some time on the pre-production service. AlastairD notes that GridFTP could be over for external transfers. With XrootD and redirection this might be solvable. Francisco asks whether it is just GridFTP or is it just XrootD? Today it is XrootD 3.3.6. Not sure when they plan to move to version 4.
9. AndreaS. ETF - mods to determine IPv6 status of services.
Done. See slides later.
10. RajaN, CostinG, AndreaS. Work with Alastair on his document.
Ongoing. AndreaS has started.
11. DuncanR. check and update the Dualstack perfSONAR mesh
Done. INFN and Edinburgh were added and some old ones were removed. It was then noted that he should not have removed FNAL. DucanR will put this back.
12. EdoardoM. check how best to set the IPv6-ready flag for VMs.
Ongoing. No progress.
13. EdoardoM. check/decide what can be published re IPv4 exhaustion at C
Ongoing.
Roundtable updates
1. FZU. JiriC reports that they are running dual-stack XrootD servers (v4.2). They observe headnode traffic but no traffic to other nodes. The head node seems to think it is IPv4 only. It is suggested that he asks the question to the mail list.
2. QMUL. Dan T report that the ATLAS CERN pilot factories now look to be safe for dual stack CEs with an IPV6 DNS entry (atlas pilot factories have ipv6 turned off to force ipv4-only behaviour). Checked for one of the (QMUL) CEs, will enable the rest next week. There is also a development IPV6 pilot factory to get ipv6 working properly.
3. RAL. TijuI report on his investigations of APEL accounting. All depends on this version of the message broker. Tried Rapid MQ and this seems to work OK.
Testing plans
We discussed the DNS Round-Robin problems whereby FTS transfers using the alias always get directed to the same node. We need to advise VO people what to do. FTS3 at CERN has 6 servers behind the alias. The ticket related to this problem has been closed without any comment as to why this was behaving like this (thought to be something different). AlastairD asks whether we have to test all load-balancing scenarios and if so what tests do we need to do? TerryF says he will suggest something.
We agree to submit an abstract to CHEP2016 (deadline 11th April). The thrust will be the experiment driven IPv6-only worker node plans. It is agreed that Alastair would be a good presenter (if he is able to attend CHEP).
DuncanR and AndreaS show their slides on progress on ETF testing (see slides on the agenda). They have used the dual-stack ETF instance at CERN and used simulated IPv6-only worker nodes (work by MarianB). MarianB comments that ETF has its own HT-Condor for submitting the test. See the results slide.
It was noted that it would be useful to invite XrootD participation in the May F2F meeting.
Ongoing Actions
1. AlastairD. Work on IPv6-only document together with other experiment reps.
2. AlastairD. Work on testing IPv6-only WN.
3. DaveK. Distribute draft IPv6 security guidance for more feedback.
4. BrunoH. Update LHCONE IPv6 wiki.
5. UlfT. Take on maintenance of checking IGTF CA CRL status re IPv6.
6. RajaN, CostinG, AndreaS. Work with Alastair on his document.
7. EdoardoM. check how best to set the IPv6-ready flag for VMs.
8. EdoardoM. check/decide what can be published re IPv4 exhaustion at C
Dates of future meetings
Vidyo conference 14 April 2016 at 16:00 CEST.
F2F meeting on 18/19 May 2016 at CERN (lunchtime to lunchtime).
Pre-GDB IPv6 workshop and GDB meeting at CERN - 7/8 June 2016.
Notes by Dave Kelsey
13 April 2016
There are minutes attached to this event.
Show them.
