SUPPORT: Runcontrol does not work for certain PVs

[John-Holt-Tessella]

Run control has stopped working for methane temperature on sandals. If you change the limits the DAE does not pause, other run control settings do generate a pause.

Acceptance Criteria

1. A PV outside the limits of run control cause the DAE to go into a waiting state
2. With multiple PVs any one of those PVs causes the DAE to go into a waiting state
3. Fix is released to all full instruments except Rikenfe and Muonfe (they don't use run control)
4. There is a test which tests run control

Notes

The only part of this ticket left is to write the test. As below it should be written against 2 blocks one from simple IOC and the Methane temperature. The test with Methane temperature should be skipped if the PV does not exist. There should also be a test for two blocks pointing at the same PV. There should be a test for two blocks pointing at the sam PV which are identical apart from the last character.

[John-Holt-Tessella]

This has been reproduced on DEMO using the blocks:

For Meth2 changing the bounds of run control makes the cross appear but the DAE doe not stop collecting data. For MethanTemp it does the correct thing.

The cause appears to be record(waveform, "$(PV):RC:_OUT") which can not read the name from the PV. It is generating IN:DEMO:CS:SB:MethanTemp:RC:_IN 2019-06-05 12:02:33.933869 LINK INVALID on a camonitor. This then is blank when adding it to the set of out of range pvs.

This seems to have been introduced in ISISComputingGroup/EPICS-RunControl@4f09172

[FreddieAkeroyd]

Can you confirm which block works - the above says"MathanTemp" works, but it seems to be the block that is mentioned as having an invalid link? Is it the name of the block or something about the order/number of blocks that causes the issue?

[John-Holt-Tessella]

Release procedure:

1. Copy file from "kits$\CompGroup\ICP\Releases\TicketXXX_runcontrol_fix\ioc\RUNCTRL\db" to "EPICS\ioc\master\RUNCTRL\db"
2. Restart the run control
3. Note change on IBEX page

[FreddieAkeroyd]

Is this only an issue for blocks pointing at PVs such as moderator temperatures? I think they do not have the long string NAME$ field as the IOC there is old.

[John-Holt-Tessella]

Maybe I will have a look.

[FreddieAkeroyd]

With regard to Meth2 and MethanTemp blocks above, if two blocks point at the same PV then you may get funny behaviour as the underlying PV name is used as the key.