Present: Fernando L, Fernando V, Bobby, Carmen, Jim, Manuel, Helfried, Fransisca, Peter, lennart, Stefan, Piotr, Clara, Wayne
1) Web Publishing of PVSS Data
Manuel gave an overview of the information collected from the experiments related to existing solutions with their limitations as well as the requirements of each. His slides can be found under:
http://indico.cern.ch/conferenceDisplay.py?confId=32790
The following major questions/points were raised:
CMS approach:
- As well as the FSM states mentioned in the slides, the Pmon states are also published. In addition, there is the possibility to access archived data as well.
- The CMS data is read only.
- Will the FSM states archived for CMS? Yes, they will be.
- Do all users have the same entry point (URL)? Yes.
- Could the FSM data not be read from the database? Yes, but then one would lose the caching mechanism and hence increase the load on the dB server.
ATLAS approach:
- Publishing of panels is not dynamic, i.e. there is a set list of panels which are updated every minute whether used or not.
- What percentage of the 50 panels is actually viewed? Not known.
- Helfried estimates that eventually about 1000 panels will be required.
- To open all the 50 panels in the 4 UIs on the single '4' processor machine takes 2-3 minutes.
- It is also possible to access the PVSS alarms on a per sub-detector basis. This uses the standard PVSS http library functions and there is no alarm reduction available.
Common requirements:
- Display of PVSS panels
- Navigate through them
- Small set regularly
- The rest on demand
- No commands to be sent
- Users: ~100
- Refresh: <60 seconds
- Protect the on-line system
There was a long discussion on how to achieve the above and it was clear that there is no obvious method. All current possibilities have problems. However, it was agree to follow up with ETM regarding the Java UI and Frank will investigate the performance of loading panels on request.
2) Plotting of PVSS Archive Data
Jim presented the two most widely used data plotting tools in ATLAS. The PDV and ROOT tools, including a demo of the ROOT tool. His slides can be found under:
His slides can be found under:
http://indico.cern.ch/conferenceDisplay.py?confId=32790
The two tools have pros and cons. The PDV can be loaded and run from anywhere but allows direct access to the dB and potentially could overload the database server. The ROOT tool requires ROOT to be installed but gives access to the ROOT canvas and hence potentially all the functionality of ROOT. After some discussion on these two approaches there was feeling that a server-based approach would be beneficial to control the access to the database and hence provide protection. Peter will present the ALICE server-based approach to accessing PVSS archived data at a future FWWG meeting.
3) A.o.B
It was decided to dedicate the next meeting on the 14th May to issues related to FSM. Clara will present the new features of the FSM component and then there will be a discussion on desired new features.
Helfried asked about how the primary gas system should publish its information from its LabView system. Wayne replied that he thought that a LabView DIP interface existed. He confirmed this following the meeting.
There are minutes attached to this event.
Show them.