I have been calling the 500/503/404 Errors "infrastructure problems" (in case there were some internal connectivity or bandwidth issues at present), but Unixchick's post from just before the latest major blockage more or less matches what I was thinking -- that would explain the seemingly excessive periods of "no new tasks" when we aren't seeing lots of HTTP errors, and might also contribute to worsening of non-BOINC services if/when they are working on something...

For instance: from somewhere around 16:30 UTC access to the non-BOINC services seemed to be a bit worse than usual, and from 18:00 UTC to 21:30 or so forum access (and use of the newer API) was more or less impossible...

By 22:00 things seemed to be more or less trouble-free again -- I wonder if they were working on something during those previous 5 or 6 hours, which would have been during their working afternoon...

Pure speculation, of course :-)

Cheers - Al.

I have had a fairly steady flow of MCM1 work over the last couple of days; in fact I've not seen any of those long intervals with only retries available, so that might be promising! However, I've seen no ARP1 work since about 21:30 UTC on 2025-07-05, and the last one of those was a retry!

I expect periods of ARP1 drought as there is likely to be a lot of supplied work still out there on systems that will take 4, 5 or 6 days to return results (if they return at all); the supply of new work doesn't seem to be a constant, smooth, process at present, so those who send work back in less than two days eventually end up on an ARP-free diet...

I may be doing better with MCM1 because I have a monitor script running that prompts for new work if it sees my relatively small buffers running low. It checks every 10 or 15 minutes and only asks the client if there's a shortfall -- this seems to be enough to stop the client from getting into one of those states where it keeps increasing the time between requests if it can't get work... Sadly, this doesn't help with ARP1 :-(

Cheers - Al.

They have completely stopped ARP WUs. even resends. I have some ARP WUs with "waiting to be sent" status as it needs another wingman. There is no official word on the cause of the complete stoppage of ARP.

MCM is flowing well at this time.

My best guess is that they are having hardware issues, and MCM is a lighter load on the system than ARP.

edit: it has been over 24 hours since even a resend for ARP has been sent out.

Quoting my self from June 28:
"Hi!
Just got several wus from a "forgotten"/left behind batches, it's no resends _0 and_1.
Batch numbers, MCM1_0235129 to MCM1_0235468 so far, none finished yet so info of cances type investgated!

Anyone having more info? 🤓

Hans S."

The "leftbehinds" for MCM1 wus have reached (What I've got at least!)

237???
238???
is the batches I've seen the last 3 days!

Maybe they are waiting for current released ARP1 wus to be returned before they start giving us Extremes and Accelerated ???🤞🏻😏

PS.
And while writing this a new update for ARP1 project appeared!

Hans S.

Extremely many "tasks are committed to other platforms" now. Very hard to get any MCM work at all.

Edit, added: Finally, between 2025-07-10 0:51:10 UTC, and 2025-07-10 0:55:25 UTC, I did get 9 new tasks. The "tasks are committed to other platforms" reply, seems to come more and more frequent.