Minutes SLAC-ALTAS TDAQ Meeting 17-Aug-2006 ------------------------------------------- Present: Stephen (phone), Sarah (phone), Steffen, Matthias, Amedeo, Su Dong, Rainer Stephen: DetCommon 0.0.1 now installed and announced. People working with the packages it includes are asked to test. Not heard anything back yet. 12.3.0 is getting built. Some packages still in there that are in DetCommon also. Will sort this out and make sure that nobody is affected by it. There is a plan to build a specific release for the Large Scale Test to be called 12.3.0-LST. The nightly tests: Looking into the way regression tests are set up for ATLAS. Current scripts may not exactly be what one wants. Taking BaBar regression tests as a possible example. Release 12.2.0 just started, so aim to import that to SLAC and try to add DetCommon to it. Sarah: Met with Richard. Plans for a mid-scale test in September that is expected to use DbProxy. Will meet with Haimo, who wants to make sure that the proxy will run. Hans, Doris and Sarah meet next week to talk about DbProxy in the Large Scale Test (no word on whether that would be canceled/postponed). Raised an nteresting issue on how the Conditions database updates get distributed to centers. Will there still be databases that might benefit from DbProxy technology? Some discussion starting on this. Amedeo: DbProxy is ready. Want people to use it as much as possible. Summarizing next issues: a) Communicating to RunControl on whether the connect failed: This should be straight-forward; b) Oracle support: Looked at SQLRelay. Looks big. ("Looks like what Frontier should have been...") Provides libraries for server and client side interfaces, including Oracle. Brings in a lot of dependencies. Will try it. There is the possibility to use it for the translator. Will check it out and look at performance; c) Multicast: Use the same test setup that we had before; d) Reliable multicast: As discussed before; e) Make a test setup in which we can start n clients (using multicast); f) Shared memory: Should follow up on the ideas raised in the meeting with Andrea. This almost certainly looks like a project highly relevant to DbProxy, so we might want to take an initiative on that. Steffen: Want to have the new BaBar farm in by Oct 20. Will keep with the plan of using the farm in IR2 until it can be moved to building 50 entirely. Su Dong: Bad news on the switch: Was about to order, but CERN IT says that with that this contract the switch must be used at CERN. Looks like we might not be able to get the CERN price but have to buy it from here. Still, purchase should proceed as soon as possible so it stays in the FY06 budget. Please post corrections/comments as a follow up.