Ok, so for the program alarm, that is alarm code 114. You can check that with V05 N09E and then you can reset the program alarm with the RSET button on the DSKY. Alarm 114 is "more marks made than desired". That alarm should only happen in P22, I think, maybe you did too many marks during a P22 earlier? When you press the mark button in P00 or so then it tends to give alarm 122 "marking not called for", which is different. Both alarms types are harmless and shouldn't have anything to do with the stuck computer activity light.
Ok, small lecture on how P00 works. Every once in a while (that time is variable) the CMC checks the current time of the CSM state vector. When that state vector is outdated it integrates both the CSM and the LM state vector to present time. Now, a few hours earlier you did V44, which told the CMC that the LM is on the surface. This inhibits the integration of the LM state vector, as that would be pointless, the LM isn't coasting in space anymore. And then, a while before LM liftoff, the flight plan calls for V45, which reset the lunar surface flag again and allows the integration of the LM state vector. But that means the LM state vector hasn't been brought to present time since the time of landing, so it's many hours old. And this is what takes the CMC so long to process, it tries to take the 20 hour old LM state vector to present time. The coasting integration routine is slow. I'm pretty sure it's a coincidence that the COMP ACTY light went on at about the same time as the LGC uplink. You simply did the V45 not long ago and then the next time the CSM state vector is outdated it will take both to present time, which takes forever for the LM one.
And here is how that went down during the actual mission:
122:31:43 Collins: Yes sir. Keep it that way. (Long Pause) Columbia is coming up on a Verb 45, Enter to reset the surface flag.
122:32:15 Evans: Columbia, Houston. Negative. Standby on the Verb 45.
122:32:52 Evans: Columbia. Roger. We copy. (Long Pause) Columbia, Houston. We've got a couple more vectors to send up to you. They'll be coming up shortly and then you can do the Verb 45 after you get those in. Over.
So they uplinked a new LM (and CSM) state vector before Collins did the V45. As that is a predicted LM state vector with a time tag after insertion of the LM the very long calculation in the CMC is prevented.
And for NASSP, I will check for the timing of the V45 and the uplinks to the CMC. The uplinks should really come first. I'll adjust the checklists accordingly.
What you can always do is a V96 which stops the P00 integration. Just do that and then let it do the uplinks of the state vectors to the CMC. That should prevent any problems.