LHC@home

... I can't run all monitoring applications.

Thanks Ivan.

Ramp-up is in progress.
Looks fine so far.

There's a problem with the JobAccountant component of our WMAgent. I'm seeing discrepancies between the job graphs and the WMStats figures. At the moment, I trust neither of them. A plea for help has been sent...

[This is internal CMS accounting, it shouldn't affect your BOINC credits.]

---

There's a problem with the JobAccountant component of our WMAgent.

I am wondering anyway why they fiddle around with the WMAgent so frequently. And still it doesn't seem to work the way it's supposed to so many times :-(

Cern-IT do the work they can do,
sometime is a IT-Problem more than you see at the beginning.
Ivan know more details about it.

Today is a greater upgrade see the News-folder.

Today is a greater upgrade see the News-folder.

thanks, maeax, for the information.
I should become better at looking up the News-folder more regularly :-)

There's a problem with the JobAccountant component of our WMAgent.

I am wondering anyway why they fiddle around with the WMAgent so frequently. And still it doesn't seem to work the way it's supposed to so many times :-(

One of the problems is that we are still using the testbed version of the WMCore applications, so we get hit more often by, well, test upgrades. The production version is updated much less often and with versions that we have already tested. We're slowly moving towards being incorporated into the production environment but progress is slow and the massive disruptions this year (and more to come, I fear) are not helping at all.
On the home front, I've just been notified that my new ADSL modem is being dispatched, so I might have a reasonably decent (10 Mbps?) connection at home for the first time in 17 months. Probably just in time for London to go into full Tier 3 lockdown...

---

... London to go into full Tier 3 lockdown...

:-( :-( :-(

Ivan, the task queue is empty. What's the reason?

Ivan, the task queue is empty. What's the reason?

I guess I lost track of when the new batch had to be submitted. More coming now...

---

Ivan, the task queue is empty. What's the reason?

All the jobs are erroring out currently (207 (0x000000CF) EXIT_NO_SUB_TASKS), so it is good that the task queue is empty.

[edit] Ivan was faster than me ;-)

---

the server status page shows a rapidly declining number of unsent CMS tasks.

Further, new Theory tasks have been unavailable for hours.

Is there a major problem somewhere?

Looking at my computers there might be something wrong with the FLOPS estimates? I have around 2800 tasks queued across my 6 computers, I doubt they can really make all of them before the deadline?

there are no new CMS tasks in the queue :-(

there are no new CMS tasks in the queue :-(

I just realized that the reason for no new tasks probably is the lack of jobs. So the automatic stop of sending out tasks in the case of no jobs works.

BOINC says "no tasks available for CMS"

I usually only run the VB's over at -dev but decided to see if I could get a CMS here and it sent me one.....lol so now I have to run it.

I got one at 28 Apr 2021, 5:06:19 UTC, could be we are just going along close to 0.

As I said before the FLOPS estimate seem wrong 2 of my computers have around 900 WU's in queue with work buffer of 0.5 day, I don't see how they can realistically do them before the deadline expires so I chew up plenty of server load for nothing. The flops estimates for my computers don't seem wrong.

no CMS tasks available for download :-(

what's the reason?

no CMS tasks available for download :-(

what's the reason?

okay, I think I found out now: my currently running WUs error out with:

207 (0x000000CF) EXIT_NO_SUB_TASKS

so the problem we had about a week ago again came up. Why that often?

no new tasks available now - is there a major problem ?