Choose timezone
Your profile timezone:
ADC Weekly
---> Rucio commissioning
rucio ping USATLAS APFs? waiting for news from them, asked just today.
next week means after Easter's Monday.
---> Central Services: OpenSSL vulnerability
Alexey Buzykaev
no questions.
---> Savannah to JIRA migration
- Dario: I'm doing things to myself. I'll create a parallel project in JIRA and only after I'll face out savannah.
- slide7 make sense! reasonable defaults should be discussed with JIRA team since now
- we need to have only ATLAS active members being able to see all the tickets, and this could be a reasonable default. Eric confirmed that it's OK for all the ATLAS computing projects
---> Software installation with optimized setup
Eric: how can happen that it took so much time to spot it before noticing it?
G: we spot immediately!
Rod: what's left we can use this? Change in the pilot with some flag in default to have the "fast".
G: we can test now with this options, then we can discuss in the second period the changes in the pilots
Rod: we are wasting a lot of time in setting up (up to 30sec) even from CVMFS.
G: minor point: the compiler is not set up by the script.
R: we need to say "this is the setup", then we have the SW team doing the validation as usual.
Emmanouil: will it be useful also for analysis? make flag with CMT, if you can restrict it would be good.
Andrej: it will be done by Paul in next pilot release...
---> Overload of frontier lpad by MC Overlay
Rod: big unbalance of how many sites are used. CERN is lowly balanced and there is a reason for this, during DataTaking it should be prevented.
Wolfgang: US is performing much better than the rest. 10M events distributed over 5 clouds, I was not expecting this load. we have improved in terms of distribution, running in 5 clouds. we are not focusing on few clouds. Special DB release will not fit here, too big. We need something on the SW side: Overlay is high priority.
in 2 weeks from now Wolfgang will try to follow up the various points.
---> Train production with NTUP_COMMON as input
Tomas: MC do you know when the updated number will be available? will know in another week.
---> AOB:
Emmanoueli : OverlayMC jobs: from the site point of view we can't do anything.
correct, GDP need to be careful. Andreu will follow up and communicate. Tomorrow's Frontier meeting will discuss the differences between the various Frontier Servers.
rucio ping USATLAS APFs? waiting for news from them, asked just today.
next week means after Easter's Monday.
---> Central Services: OpenSSL vulnerability
Alexey Buzykaev
no questions.
---> Savannah to JIRA migration
- Dario: I'm doing things to myself. I'll create a parallel project in JIRA and only after I'll face out savannah.
- slide7 make sense! reasonable defaults should be discussed with JIRA team since now
- we need to have only ATLAS active members being able to see all the tickets, and this could be a reasonable default. Eric confirmed that it's OK for all the ATLAS computing projects
---> Software installation with optimized setup
Eric: how can happen that it took so much time to spot it before noticing it?
G: we spot immediately!
Rod: what's left we can use this? Change in the pilot with some flag in default to have the "fast".
G: we can test now with this options, then we can discuss in the second period the changes in the pilots
Rod: we are wasting a lot of time in setting up (up to 30sec) even from CVMFS.
G: minor point: the compiler is not set up by the script.
R: we need to say "this is the setup", then we have the SW team doing the validation as usual.
Emmanouil: will it be useful also for analysis? make flag with CMT, if you can restrict it would be good.
Andrej: it will be done by Paul in next pilot release...
---> Overload of frontier lpad by MC Overlay
Rod: big unbalance of how many sites are used. CERN is lowly balanced and there is a reason for this, during DataTaking it should be prevented.
Wolfgang: US is performing much better than the rest. 10M events distributed over 5 clouds, I was not expecting this load. we have improved in terms of distribution, running in 5 clouds. we are not focusing on few clouds. Special DB release will not fit here, too big. We need something on the SW side: Overlay is high priority.
in 2 weeks from now Wolfgang will try to follow up the various points.
---> Train production with NTUP_COMMON as input
Tomas: MC do you know when the updated number will be available? will know in another week.
---> AOB:
Emmanoueli : OverlayMC jobs: from the site point of view we can't do anything.
correct, GDP need to be careful. Andreu will follow up and communicate. Tomorrow's Frontier meeting will discuss the differences between the various Frontier Servers.
There are minutes attached to this event.
Show them.