JCOP FWWG Meeting 20220222
INDICO: https://indico.cern.ch/event/1127482/
News on JCOP Framework and WinCC OA
- fwInstallation tool ver. 8.4.5 will be released today, notably with fixes related to the use in redundant systems, requested by CMS
- import-checker ver 0.0.1 released with features requested by users after its first release
- patch P007 for WinCC OA 3.18 - will not be evaluated
- new policy for Microsoft products at CERN: support for terminal server installations; possible changes for groups/egroups
Presentation of ATLAS OPC-UA server for ATCA Shelf Manager
- see the slides attached, presented by Paris
- a few micro-TCAs already installed in CMS, with more to come in LS3, also ATCAs. Not yet sure what will belong to DCS what to the DAQ and where the control will be done. Great to know that there is a solution existing...
- In ATLAS there was a similar uncertainty, yet auto-discovery and integration of extra functionality (eg readings from embedded FPGA fed to FSM) convinced the TDAQ community (due to all the integrations, archiving, etc)
- Generation and templates: seems generic. Any plans to separate it into a dedicated module? The idea is good, but the information used by Jinja is strictly related to OPC-UA... Hence could not be reused fully genericaly. But the concept certainly reusable. Quasar itself also uses a lot of Jinja.
- Use for monitoring of network or server equipment? Should not be difficult to do it - not sure about interest in it - server hardware monitoring done already differently by the sysadmin teams using datacenter tools.
OPC-UA servers for ATCA Power Sources - update on the status
- See Ben's slide
- invitation to REVIEW THE SPECIFICATION and comment on it => link in the slide
- ATLAS subdetectors have pieces of this hardware already in P1, and currently switch it on/off manually. LS2 is still low-scale hence capable; for LS3 will be in high number, hence more and more critical; the sooner the better
WinCC OA and non-reachable OPC UA server items - current status
- See Mariusz's slides
- topic discussed already in the FWWG meeting in January; providing an update on current understanding of the problem from the experts point of view
- comment from Piotr Nikiel: BadNoCommError on start -> a feature of Quasar should be tested.
- comment from CMS: the question of compatibility needs to be addressed by subdetectors - their code may take actions, etc, and see if this is what they want. It would be good to have a flexible mechanism: keep BadWaitingForInitialData and have a config file entry determining if it should be transmitted to WinCC OA or not, yet this is not the full solution (for the problem of restarting the server and hw not being available).
- CMS: if you restart canOpen OPC server - what is the status of ELMBs that are not reachable? Some things will be BadForInitialData which is quite natural. The situation is not really well detected (no replies from ELMB). In the new OPC Server (NG) it will change: BadWaitingForInitialData will be a transitional state.
- PiotrN: OPC UA does things so much cleaner than the DA... one should not really model the behaviour on the DA...
- Stefan: default behaviour should be that the status is propagated maybe with an option to ignore this. While waiting for "a fix", one could change different subscription for e.g. old and new CAEN crates and adjust the behaviour (similarly to what ELMB already does).
- Mariusz: Modification of subscription setting of fwCaen: not really supported in the device model to support different statuses for different crates and boards (depending on the type of board). But this could be set up by the subdetector expert
- PNikiel - ATLAS has a number of tools that may be used for investigation and testing
- PNikiel - OPCUA endpoints are open to early... maybe open them later on, once data is available... maybe a feature to be added to Quasar?
- Having BadWaitingForInitialData as the initial state is a conscient choice of the server's author. Quasar is not even specifying this - the code author specifies it. But this value seems to be the right choice!
- Cristina:A nyway, a change in the OPC-UA startup sequence from slide 8 could also be implemented and seems to be delivering what is needed
- Valeri: NA62 is good with the proposal of a short transition in the OPC Server. Invalid bit is not an issue.
- Clara: LHCb: trust in good sense! What is certainly not good is not to know about the problem
- Mateusz: ALICE - similarly. Also, recently observed problems with sending commands to Wiener and failures not being visibe in any way -> similar problem to those reported by CMS and ATLAS already...
- Agreed to continue further discussions offline, and report back during the next 2 weeks at the FWWG.
- Frank: the off-line discussion should also consider concrete use-cases from the experiments, not only the s/w-design perspective.
There are minutes attached to this event.
Show them.