For information regarding ARP1...

On the evening of 2025-08-07 Dylan posted the following in a No Tasks thread in the BOINC Agent support forum just after a user post about lack of ARP1:
Which leaves the question of whether things are fixed yet, despite the reappearance of MCM1 work!...

MCM1 workflow restarted some time before 22:00 on 2025-08-08 (that being when one of my systems started loading up again) -- all of my systems were up to their quotas by about midnight UTC and are keeping topped up when they can be bothered to ask for work... Almost all the work received has been new ; just the occasional retry for a download error. That looks promising :-), though things may change if Dylan has more fixing work to do today.

If the MCM1 issue has been solved, it will be interesting to see how long it will be before we get another drought (if at all). At the moment, I doubt there are any late return retries left over (especially given that just over 40% of the tasks I got over the last three days were missed deadline retries, and I guess others were seeing the same sort of level if they were getting work at all!).

However, if all the problems have been resolved, I wonder why the ARP1 work creation hasn't been restarted...

Cheers - Al.

We had a nice weekend of all you can crunch MCM.

I'm really hoping they start up ARP again soon. I didn't get any resends over the weekend, but the reference number means there are still some resends bouncing around.

Another round of MAM beta would help too. The sooner the linux, and windows people get MAM out of beta and are really crunching it, then we find out if they are going to let Darwin have a piece of it too.

I had to turn MCM1 back on due to the ARP1 issue. Yep, getting lots of those, but zero ARP1 new or re-send. Oh well.

Yup, I saw my last new MCM1 tasks just after 04:30 UTC today, and collected a single retry (for a Detached task) a few minutes later. Nothing since.

This was about due to happen if the usual pattern from before the "disc full" induced problems was going to resume -- it's about the time when the first missed deadline retries would be going out, and we seemed to get a prolonged retries-only phase around then (recently several were well over 24 hours rather than the odd 6 hours or so we had been seeing for months before...)

For what it's worth, there still seem to be a few ARP1 retries around; I got one yesterday :-) and keep seeing that "other platforms " message in a context that means it's about ARP1 as well as the usual "no tasks are available"...

By the way, there seems to be another issue. I monitor the transition of my returned work from validation to database purge and yesterday morning I noted that there didn't seem to be much if any stuff making its way through to the "ready to delete files" stage. This morning the only things going through seem to be the very small number of ARP1 tasks that have validated over the last few days, so it looks as if all the MCM1 assimilators have gone missing, possible as far back as 36 hours ago. It's either that or someone has decided that my work shouldn't get fully signed off :-)

However. validations seem to be happening o.k. at present (not that that's going to matter much soon!) -- my number of Pending Validation tasks is staying fairly steady but the number of validated tasks awaiting assimilation is going up rapidly. If all the remaining pending tasks validate before assimilation restarts I'll have nearly 3500 results stacked up (as against the more usual 800 to 1400, depending on validation rates)

I just hope we aren't about to get a repeat of the late 2023 to Spring 2024 assimilator problems...

Cheers - Al.

P.S. the full analysis above is about an hour behind time, but a quick check just before posting suggests no change...

Seven _2 resends, 2025-08-13 22:16:43 UTC. That's all since the last "new" MCM task 2025-08-13 03:55:01 UTC. Total silence from the WCG team.