SLAC/ATLAS TDAQ Meeting, 18 Jan 2007 Present: Sarah, Steffen, Amedeo, Rainer, Su Dong Sarah: No news from Radovan. Were expecting to hear from him Mon/Tue of this week. Sarah already pinged him. Rainer will send a message to ask about an ETA. Investigate if replacing the MySQL library in CORAL with SQLRelay is an option. SQLRelay is supposed to be usable as a drop-in library. Hans and Richard asking us to investigate this before approaching the CORAL people. CORAL would support SQLRelay, the proxy would talk SQLRelay. In this case a MySQL connection would also go via SQLRelay. No change to HLT, but we would need to implement SQLRelay as our internal protocol. It may even be a run-time substitute, so we should be able to test this ourselves. Farm tests: Righ now editing the xml files by hand. Rainer tested from a different userid and encountered some ownership/filename issues. Will get path names straightened out. Replace controllers with dbstressor controllers. LST report is scheduled for Feb 8. Sarah is helping to put this together. May be insteresting to look at Serge's and Haimo's reports. Steffen: Farm will have to move. Probablay from 2nd to 1st floor. This to be schedule within the next few months. The interruption should not be more than a few days. Do we need the big switch and the second (optical) GBit ethernet? Discussion if that could that be switched off while not in use. We certainly want to keep the option. There is a danger that the line cards inside, which are official SCCS spares, might vaporize if powered off. Right now only using it as a router. Conclude to agree to switching it off, but make sure that it stays fully cabled, will be moved etc. Use the router capabilities of the ProCurve. May be acceptable. One line card is actually still in IR2. Amedeo: Plan: Instrument the proxy. Get dbproxy timing information out of the farm. Learn to run the farm in a reliable way and be able to do all tests. Instrumentation can be done in parallel. Then try out the SQLRelay approach.