Choose timezone
Your profile timezone:
Powered by Indico
v3.3.3-pre
Attended:
Alessandra ( Forti), Alessandro (Paolini), Andew , Balasz, George, Jarka, Kyle,Oxana, Pablo,Vincenzo, Alberto, Alexey, Salvatore, Maarten, Giuseppe, Stephan, Marian, Julia
Discussion after presentation of Andrew:
How often the GocDB update should happen? Once per day.
Maarten expressed some concern that there can be dangerous to allow automatic update of GocDB which was always a stable repository of the GRID resources and sites. It can be polluted by mistake. After some discussion, there was a conclusion, that in case new services are not introduced through the automatic update and nothing can be deleted through the automatic update, there is no much danger. Julia asked Andrew, whether Dirac system would query GocDB for static information rather than BDII in case all necessary static information is being in GocDB, Andrew replayed that yes, this is the plan, but reading from GocDB has to be in place. One of the possibilities is to populate GocDB querying central BDII. Special privileges will be required for this purpose. Overall, reported experience is very positive and follows the direction we would like to go.
Discussion after Pablo's presentation:
Everything which is required is in place and is enabled for ATLAS and CMS. Can be done also for ALICE and LHCb on their request. Many thanks , Pablo. This new type of metric (production/non production) which has impact on the availability calculation algorithm, can be used for other purposes, for example to enable condition "For T1 sites only" by introducing a new metric like T1 (yes/no).
Production/ non production metric has been introduced on CMS request, CMS is happy that the possibility to test services without taking results of tests of services not being in production into consideration is implemented.
After the meeting , Stefan asked by mail , whether it is fine to introduce in the VOfeed an attribute for VOfeed version. Need to check with the task force members, in particular with Pablo and Marian to check that it does not break anything. Action: to check before the next meeting.
Status of migration of responsibilities for VOfeed
CMS: Done. It is already running on the CMS node. The only problem is that in order to make a redirection, the old machine is still used. Pablo is checking whether it can be avoided.
ALICE : Not yet
Next meeting is on the 23rd of February.
Main topics are :
1).Storage object discussion taking a concrete example of the KIT storage and LHCb and ATLAS use cases
2). How we ensure common naming convention for the service types along the whole chain of our systems (GocDB/OIM, VOFeed, ETF, SAM, CRIC, experiment specific systems). OSG colleagues confirmed that they will join the meeting.