Using epics7 env for rocky9 compatibility

[KaushikMalapati]

Description

motor-expert-screen was using epicsenv-3.14.12.sh, but the oldest version of epics base available on rocky9 is 7.0.2, so motor-expert-screen will use that environment from now on

I also had to update how :PN motors are opened, because originally they executed a file that tried to activate the epics 3.14 environment. Now I am directly opening the screen with edm and I added the path to EDMDATAFILES

Motivation and Context

https://jira.slac.stanford.edu/browse/ECS-6842

How Has This Been Tested?

Interactively

Where Has This Been Documented?

n/a

[ZLLentz]

I agree with this conceptually and I agree with the implementation
Let me double-check a few screen loads before/after on a hutch machine to make sure there aren't any obvious regressions

[ZLLentz]

This PR absolutely enables this script to be run from rocky9 hosts! Very good.

Here are the PV bases I'm planning to compare:

IMS: XPP:USR:MMS:01
PP: XPP:SB2:MMS:29
Newport: XPP:USR:MMN:01
AeroTech: XCS:SND:T1:Y1
SmarAct: XPP:MCS2:01:m1
Beckhoff (old): MFX:LJH:JET:X
Beckhoff (new): IM3L0:PPM:MMS

So far, the IMS was the exactly the same but the PP (old IMS) picked a slightly different screen (top is with the PR, bottom is before this PR):

I'll continue through the rest, it shouldn't take too long

[ZLLentz]

Everything was the same except for these super old pulse picker screens which are now using a less readable variant

[KaushikMalapati]

I will move the update EDMDATAFILES to use the old pcds_motionScreens directory. I assumed the newer version would just be better, but I guess not. Thanks for checking all this. Do you think there is anything else I should test?

[ZLLentz]

I think if for a bunch of PVs and on the hutch operator machines (-daq, -control) we get the same results, then everything should be good.

We want to enable this on rocky9 but not at the expense of ongoing experiments who may rely on this script.

[KaushikMalapati]

I have tested this on these machines

• psbuild-rhel7
• cxi-hutch01 (rocky9)
• ioc-mfx-misc01 (rhel5)
• rix-monitor
• tmo daq
• txi-console
• mfx-console
• cxi-control

And everything looks identical to me. The only "problem" I saw was that on ioc-mfx-misc01, the typhos screens did not open. When I tried to run motor-typhos directly (motor-expert-screen redirects stderr/out to /dev/null) I saw these errors

/cds/group/pcds/pyps/conda/py39/bin/python: /lib64/libc.so.6: version `GLIBC_2.10' not found (required by /cds/group/pcds/pyps/conda/py39/bin/python)
/cds/group/pcds/pyps/conda/py39/bin/python: /lib64/libc.so.6: version `GLIBC_2.7' not found (required by /cds/group/pcds/pyps/conda/py39/bin/python)
/cds/group/pcds/pyps/conda/py39/bin/python: /lib64/libc.so.6: version `GLIBC_2.9' not found (required by /cds/group/pcds/pyps/conda/py39/bin/python)
/cds/group/pcds/pyps/conda/py39/bin/python: /lib64/libc.so.6: version `GLIBC_2.6' not found (required by /cds/group/pcds/pyps/conda/py39/bin/python)

I ran it on another rhel5 host, ioc-cxi-usr1 and got the same results. I don't think this is a problem though since nobody really opens screens from rhel5 ioc hosts. If it, I imagine that installing one of the listed versions of glibc would fix it.

I used the pvs you tried plus MEC:USR:MMC:01 for mmc and TMO:MCS2:02:m4 for unencoded smaracts. I think this is ready, but I can wait or test this some other way if you woudl like.

[ZLLentz]

I think you've done more than enough testing!
I had forgotten that the typhos screens don't open on rhel5. Maybe this is OK though (and certainly it predates this PR)