LHC@home

Our database cluster is heavily loaded today, and LHC@home services time out from time to time. Our DBA is trying to fix this. Sorry for the trouble and happy crunching.

The problem now, unfortunately, is, that downloading new tasks is impossible, since after several automatic unsuccessful trials of the BOINC manager to download tasks, communication to the download server is being deferred for 24 hours :-(((
Something should be changed with this system.

Although the server status shows "Workunits waiting for validation: 0"
my hosts have 18 CMS WUs waiting for validation since 2021-03-11.

CMS has a quorum of 1 which is already fulfilled.
Hence, I guess they will never be send out to another host to get checked.

The problem now, unfortunately, is, that downloading new tasks is impossible, since after several automatic unsuccessful trials of the BOINC manager to download tasks, communication to the download server is being deferred for 24 hours :-(((
Something should be changed with this system.

is there anything I can do to escape this strange trap, or do my machines really have to wait until tomorrow to get new tasks?

Although the server status shows "Workunits waiting for validation: 0"
my hosts have 18 CMS WUs waiting for validation since 2021-03-11.

CMS has a quorum of 1 which is already fulfilled.
Hence, I guess they will never be send out to another host to get checked.

what I just notice:
many tasks that were uploaded yesterday afternoon and evening are still waiting for validation (CMS, Theory, Sixtrack), but 2 CMS tasks that were uploaded about 1 hour ago were validated immediately.

So I am wondering whether the tasks from yesterday will ever get validated.

Why don't you just put a second server in parallel!!! Eric

---

Although the server status shows "Workunits waiting for validation: 0"
my hosts have 18 CMS WUs waiting for validation since 2021-03-11.

Same here, i have 72 sixtrack waiting for validation.
Maybe "status page" is not updated correctly.

Same here, i have 72 sixtrack waiting for validation.

Now 86.
Seems that validator have some queues to solve.

We are checking our result records and will trigger a re-validation of results that did not get validated properly during the DB outage.

The problem now, unfortunately, is, that downloading new tasks is impossible, since after several automatic unsuccessful trials of the BOINC manager to download tasks, communication to the download server is being deferred for 24 hours :-(((
Something should be changed with this system.

is there anything I can do to escape this strange trap, or do my machines really have to wait until tomorrow to get new tasks?

I thought Boinc backed off a slightly longer time each time there's no work available. That makes sense. If there's no work, try in 10 minutes, still no work, must be bigger problem, wait an hour, nothing then, wait a few hours, still nothing, wait a day, don't keep pestering. Seems sensible to me, it's what I'd do if I was trying to phone a company with busy phone lines. You can always just select the project and click update.

---

Since 5.00 UTC Atlas-Tasks are back!