Present:

• Saima, Lee, Oliver, Frank (CMS)
• Eva, Dirk (IT)

Discussion

work plan details

• Agreed to start with the first two items on the work plan now (functionality and scalability test). This first phase is expected to take two weeks and will be done in parallel to the stress tests on the validation cluster to avoid interference between both tests.
• Data model and population jobs are existing.
• Filtering of delete statements will need to be implemented during the streams setup and should be tested as part of the functionality tests

• Second part of the test plan (finishing scalability and failure mode tests) are planned for one week after the first phase. At this point integration with the offline RAC will be attempted to move to a more realistic hardware scenario.

• Tentative date for next meeting: Tue 1Nov 15:00 (CERN time)

Actions

• Eva will give access to the 3D replica as soon as the network intervention is finished (Tue afternoon).
• Lee will prepare a summary of the CMS database strategy and T1/T2 resource and networking requirements (squid box cpu and storage) for the upcoming GDB meeting
• CMS will define a experiment database contact which will collect and prioritize new account and resource requirement inside CMS.
• Saima will produce a target transfer rate estimate (online->offline) based mainly on Pixel data