Choose timezone
Your profile timezone:
ADC Weekly
---------------------->>>>>>>>>>>> Rucio migration - new stable DDM Clients 2.5.0
Ilija: localSetup EMI: I need RUCIO_ACCOUNT variable needs to be set. my cronjob were failing. Message is repeated indefinitely, needs to be fixed.
Vincent: if you've a valid proxy this would have not happened.
AleDG: we should have announced more broadly, apologies.
We need to follow up the DQ2 upgrade on all the service: e.g. PandaServer has a jira ticket open, waiting for reply since Friday.
---------------------->>>>>>>>>>>> NTUP Common
David: is it still ok to remove data v1 if not used >6months?
Nurcan: mm, no
Sabine: people have complained about availability of NTUP COMMON
Nurcan: I've mentioned it on DAST mailing list. The problems were corrupted files…
Sabine: mmm, no, it was more on availability… it's on the top convener.
Andrej: I'll make my question offline.
---------------------->>>>>>>>>>>> ATLAS SAM tests - a report on experiences at FR sites
AleDG downtimes "issues" should be identical to the OPS one
Sabine: actually we didn't see this for OPS… needs to be investigated
AleDG: very useful the table with numbers (with also HC). --> to be discussed/followed up
WMS timeouts needs to be enlarged (because we don't use the CE in the same way WMS use them)
---------------------->>>>>>>>>>>> CNAF - FNAL WAN access. Salvatore
KIT: sites with Alice can provide infos. we saw in the past that alice was using only xrootd federated.
Rob: slide4 is that % of total traffic? Do you have a plot which shows how much you have really on the network link
Salvatore: 400MB/s was the rate with which FNAL was reading data from CNAF
Rob: CMS is using federated access since 3 years: how often it happened?
-- zero
Salvatore: the point is that we don't want this to happen during data taking
Rob: yes…
Rob: what capacity is the Tier1 going to have?
Salvatore: it should be upgraded up to 40Gb/s.
the message is that the network is a resource which is limited…
Alaistair: (answering Rob) RAL 2/3 times a year we see CMS having network incidents!
Vincent: it's impossible to fairshare… this is the worrying part.
Ilija: there is a config in xrootd to share bandwidth between experiments. that's how we do in BNL now.
Rob: we might have something in xrootd (partitioning proxy servers) but I see the question "what about the other traffic". we need to have reasonable configuration, at least in Tier1s.
Ilija: there will be federated data storage in April, it will be a good place where to discuss with all the experiments.
---------------------->>>>>>>>>>>> Rucio Renaming:
Hiro: do we need to rename host disk?
yes
Rob: when can we start consolidating spacetokesn? Before DC14?
Cedric: no, actually not before Rucio is fully commissioned (summer)
---------------------->>>>>>>>>>>> auto-setup ALeDS: we are ready to go, but in US is not easy to debug: ALeDS will follow up with John Hover and Jose (APF in US)
---------------------->>>>>>>>>>>> Hiro: lcg-cp (lcg clients) form CVMFS crashes with certain OpenSSL version on WN
ALeDS: it's a known issue, we need to rebuild the cache, he will follow up.
---------------------->>>>>>>>>>>> Alexey Sedov:
on tapes there are new families?
David: I can explain: some where subscribed (and deleted) by error.
for the other we have to follow up, Alexey will
---------------------->>>>>>>>>>>> ADCoS report from Armen (forgotten, apologies from AleDG)
Ilija: localSetup EMI: I need RUCIO_ACCOUNT variable needs to be set. my cronjob were failing. Message is repeated indefinitely, needs to be fixed.
Vincent: if you've a valid proxy this would have not happened.
AleDG: we should have announced more broadly, apologies.
We need to follow up the DQ2 upgrade on all the service: e.g. PandaServer has a jira ticket open, waiting for reply since Friday.
---------------------->>>>>>>>>>>> NTUP Common
David: is it still ok to remove data v1 if not used >6months?
Nurcan: mm, no
Sabine: people have complained about availability of NTUP COMMON
Nurcan: I've mentioned it on DAST mailing list. The problems were corrupted files…
Sabine: mmm, no, it was more on availability… it's on the top convener.
Andrej: I'll make my question offline.
---------------------->>>>>>>>>>>> ATLAS SAM tests - a report on experiences at FR sites
AleDG downtimes "issues" should be identical to the OPS one
Sabine: actually we didn't see this for OPS… needs to be investigated
AleDG: very useful the table with numbers (with also HC). --> to be discussed/followed up
WMS timeouts needs to be enlarged (because we don't use the CE in the same way WMS use them)
---------------------->>>>>>>>>>>> CNAF - FNAL WAN access. Salvatore
KIT: sites with Alice can provide infos. we saw in the past that alice was using only xrootd federated.
Rob: slide4 is that % of total traffic? Do you have a plot which shows how much you have really on the network link
Salvatore: 400MB/s was the rate with which FNAL was reading data from CNAF
Rob: CMS is using federated access since 3 years: how often it happened?
-- zero
Salvatore: the point is that we don't want this to happen during data taking
Rob: yes…
Rob: what capacity is the Tier1 going to have?
Salvatore: it should be upgraded up to 40Gb/s.
the message is that the network is a resource which is limited…
Alaistair: (answering Rob) RAL 2/3 times a year we see CMS having network incidents!
Vincent: it's impossible to fairshare… this is the worrying part.
Ilija: there is a config in xrootd to share bandwidth between experiments. that's how we do in BNL now.
Rob: we might have something in xrootd (partitioning proxy servers) but I see the question "what about the other traffic". we need to have reasonable configuration, at least in Tier1s.
Ilija: there will be federated data storage in April, it will be a good place where to discuss with all the experiments.
---------------------->>>>>>>>>>>> Rucio Renaming:
Hiro: do we need to rename host disk?
yes
Rob: when can we start consolidating spacetokesn? Before DC14?
Cedric: no, actually not before Rucio is fully commissioned (summer)
---------------------->>>>>>>>>>>> auto-setup ALeDS: we are ready to go, but in US is not easy to debug: ALeDS will follow up with John Hover and Jose (APF in US)
---------------------->>>>>>>>>>>> Hiro: lcg-cp (lcg clients) form CVMFS crashes with certain OpenSSL version on WN
ALeDS: it's a known issue, we need to rebuild the cache, he will follow up.
---------------------->>>>>>>>>>>> Alexey Sedov:
on tapes there are new families?
David: I can explain: some where subscribed (and deleted) by error.
for the other we have to follow up, Alexey will
---------------------->>>>>>>>>>>> ADCoS report from Armen (forgotten, apologies from AleDG)
There are minutes attached to this event.
Show them.