SAM EGI transition and SAM3 UI
presents in the room: Pablo, Maite, Stefan, Marian, Luca, Lionel, Ivan, Julia, Nicolo (LAST!)
Remote: David, Pepe
Apologies: Pedro, Costin
----Review of actions from jira
All items for January closed -- good.
8 items for Feb, some of them already closed. For the open ones:
*) Pablo, "evaluate common metro store for REBUS" and is going quite well.
*) Eddie, "evaluate HC schema",
*) Julia, documentation (drupal or twiki)
*) Luca, simplification on SAM/Nagios framework
*) Ivan, downtime availability.
If there are any issues, please speak up.
Luca: simplification on sam/nagios framework: many meeting/brainstorming, and we are evaluating various options. Since we are broading the scope, it might take some more time.
Pablo: deadlines are not cast on stone, if there are good reasons we can postpone them.
We also added one more item (to be accomplished before the end of March): publish ALICE data.
----- First topic from Marian: SAM transition to EGI
Pablo: the bridge from OSG is also used to collect the downtime for WLCG?
Marian:yes, even if we decommission we need something similar, but it's a minor thing.
Maarten: will the other brokers be outside CERN?
Lionel: yes, in croatia and greece.
Maarten: things are going ok, still it seems a "tight" schedule.
Julia: Marian won't stop working on the topic if we are a bit late.
Maarten: on the technical challenge "stable operations of EGI network necessary", which are the worries? will they be able to work and maintain the broker on their own?
Lionel: yes officially, of course we will help if needed.
--- presentation finished. Q
Maite: will we be able to receive the results of the tests in the "site nagios"?
Pepe: we are working on it, for the SAM3. Instead of getting all the metrics as today we will receive an overview with links.
Maite: we run a full nagios server that collects the data from SAM/Nagios, and we have Nagios to SLS. in the future we can use other possibilities.
we have a release and we just install the new release, including probes, which we found very useful. how will it be in the future? for the service managers it is not easy to understand and debug the probes of the experiments
Maite: we need a way to really test that our instance X is not working. we use it in daily operations. The tests are different than the experiments tests, since we need to know if particular instances are broken.
….. discussion to see if the OPS tests could be included in a non-critical profile of the experiments, and if the site administrators would understand the experiment probes
Pablo: Let's continue the discussion offline, and report back in the next meeting
---- Second topic from Ivan
slide4 profiles are in synch with what we have now. still some issues. anyway you can try them out.
profiles: if you want to modify algorithm then we need to delete it.
Alessandro: if we add a new service from day x, what will happen? will the numbers be recalculated? it seems (to be verified)only the last week will be mixed..
Maarten: can we have versioned profiles?
Pablo: we will re-discuss this. For now is OK.
Julia: why sometimes the speed is so slow?
Pablo+Ivan: we have some idea, we are working on it.
next meeting will be in 2 weeks from now, 28th February
main topics HC functional testing. and the Q of Maite.
--> Nicolo' volunteered for the minutes.
There are minutes attached to this event.