LHC@home

Are you not getting any?

I got tasks during this day, but now again it says "no tasks available for CMS simulation, although the Project Status Page shows about 190 unsent.

Edit: just to find out, I tried to download a Theory task and a LHCb task, for both plenty of unsent should be available according to the Project Status Page.
However, no success either. So something seems to be wrong somewhere.

EDIT: now, all of a sudden, new CMS tasks were downloaded :-)

Strange, tho' I note that there has been a drop-off in running jobs from about 1715.
You don't have a time-of-day-controlled firewall by any chance? :-0!

---

the Project Status Page once more shows "0" unsent tasks for CMS, as well as for Theory and LHCb.
Any idea what's going on?

the Project Status Page once more shows "0" unsent tasks for CMS, as well as for Theory and LHCb.
Any idea what's going on?

No, not yet. There doesn't seem to be any change in the number of running jobs. Investigating...

---

right now, the Project Status Page shows 200 "unsent" CMS tasks (and about same number for LHCb and Theory).
However, when trying to download any, BOINC says "no tasks available"

I guess now it seems clear what the problem is - Nils wrote in the Sixtrack Thead short time ago:

"...Sadly yesterday and today the problem is much worse than usual as the response-time of our database is degraded and BOINC has problems pulling out tasks from the DB. At times BOINC clients will not be able to fetch any task from LHC@home, not even for the VM applications..."

Yes, that seems to be part of the problem. Right now I'm getting "Server error: feeder not running", and the server status hasn't been updated recently.

---

Yes, that seems to be part of the problem. ...

It's really too bad that since about Mid-December, when the huge ATLAS desaster startet (followed by all the Sixtrack problems), the whole LHC situation becomes worse and worse.

So far, CMS, LHCb and Theory were NOT affected - until day before yesterday. From then on, ALL sub-projects are having major problems :-(

I got some tasks at 1909, and the server status page is showing some tasks available. So far the number of running jobs hasn't dropped too far, but it remains to be seen if we can get back to the levels of last weekend. I need to submit a new batch of jobs tomorrow; before they take effect we might get a shot at some smaller test batches that are waiting to run, hopefully with some of them going to a new Tier-3 site we are setting up (it will be Laurence's farm eventually, but currently I believe just one VM is involved). If we get that to work we will be a lot closer to integrating into the CMS Production team upon which my need to continually monitor will drop away (just in time for retirement?).

---

early evening yesterday, new CMS tasks could be downloaded :-)

However, what I notice on my tasks list (website): for all CMS tasks that were finished and uploaded from about yesterday noon on, in the column "credit" it says "pending".
With CMS, I never saw this before. Normally, the credit shows up short time after upload.
What's wrong?

... What's wrong?

Nothing, as long as your tasks are in the validation queue.

See:
https://lhcathome.cern.ch/lhcathome/server_status.php

Task data as of 3 Feb 2018, 7:58:10 UTC
Workunits waiting for validation	298

... just in time for retirement ...

:'(

See:
https://lhcathome.cern.ch/lhcathome/server_status.php

Task data as of 3 Feb 2018, 7:58:10 UTC
Workunits waiting for validation	298

it's hard to image that only 298 tasks are waiting for validation. If you figure how many Sixtrack tasks are being uploaded permanently, and are then waiting for validation over days and even weeks; so, the figure "298" is definitely wrong.

I agree with Erich56, the server status shows now pending as 327 and I have on my 3 hosts about 650 pending validations.

---

See:
https://lhcathome.cern.ch/lhcathome/server_status.php

Task data as of 3 Feb 2018, 7:58:10 UTC
Workunits waiting for validation	298

it's hard to image that only 298 tasks are waiting for validation. If you figure how many Sixtrack tasks are being uploaded permanently, and are then waiting for validation over days and even weeks; so, the figure "298" is definitely wrong.

Well, at the end it's just a status flag in the server's database that is not yet set for whatever reason (guess due to the high load you mentioned). Nothing you should be worried about as the number is not very large nor can you do anything on your side to speed up the process.

ATM the SSP shows a Transitioner backlog of about 16 h.
So i think that is the reason for the wrong WU Status shown on the task list.
I see these false status flags also on my task list for sixtrack and Atlas.

The good news is:
Sixtrack tasks are downgrading to ZERO:
At the moment:177k.
So, if this is realy and no new Sixtrack tasks for the pipeline, we will see tomorrow a better performance for the VM-Projects :-))

The good news is:
Sixtrack tasks are downgrading to ZERO:
At the moment:177k.
So, if this is realy and no new Sixtrack tasks for the pipeline, we will see tomorrow a better performance for the VM-Projects :-))

The bad news is:
The WMAgent is having problems.
My monitor shows only about another 40 minutes before the Condor job queue is empty.
Messages sent, I hope someone at CERN is checking their e-mails.

---

Ivan,
thank you, your work seem 24/7 ;-).

The job queue is dry. Set No New Tasks to avoid BOINC errors.

---

The job queue is dry.

Confirmed.
Thanks Ivan.