IT guys sneakily moved server IP. Took us a bit to notice and now we are having some issues sending out jobs. We are hopefully gonna fix it today and send out lots of sims.

IT guys sneakily moved server IP. Took us a bit to notice and now we are having some issues sending out jobs. We are hopefully gonna fix it today and send out lots of sims.

I noticed that the number of 'tasks in progress' on the server status page dropped very dramatically, from around 2,000 this time last week to barely 100 now.

At the same time, most of the very rare WUs that I received were resends from user errors like timeouts or crashes.

So it seemed to me that the process of generating new work from previous successful runs had ground to a halt.

All BOINC projects have a 'workunit generator' of some sort, perhaps under a different name like 'splitter'. Usually, the status of these daemons is displayed on the respective projects' server status pages, but not here.

Would it be worth adding the extra daemon status line to the SSP, so we can monitor it for you?

IT guys sneakily moved server IP. Took us a bit to notice and now we are having some issues sending out jobs. We are hopefully gonna fix it today and send out lots of sims.

I noticed that the number of 'tasks in progress' on the server status page dropped very dramatically, from around 2,000 this time last week to barely 100 now.

At the same time, most of the very rare WUs that I received were resends from user errors like timeouts or crashes.

So it seemed to me that the process of generating new work from previous successful runs had ground to a halt.

All BOINC projects have a 'workunit generator' of some sort, perhaps under a different name like 'splitter'. Usually, the status of these daemons is displayed on the respective projects' server status pages, but not here.

Would it be worth adding the extra daemon status line to the SSP, so we can monitor it for you?

Hm I don't think that's feasible. The workunits are generated in a highly complex manner on our local machines using adaptive sampling methods.
Practically we just have to monitor how many WUs are running and when it drops too much we should check if our machines are still doing their jobs.

IT guys sneakily moved server IP. Took us a bit to notice and now we are having some issues sending out jobs. We are hopefully gonna fix it today and send out lots of sims.

Check the forums here. Most active projects will get a post within a day of not getting any tasks. GPUGrid should be even quicker with the nearly zero queue.

IT guys sneakily moved server IP. Took us a bit to notice and now we are having some issues sending out jobs. We are hopefully gonna fix it today and send out lots of sims.

Check the forums here. Most active projects will get a post within a day of not getting any tasks. GPUGrid should be even quicker with the nearly zero queue.

+ 1

Yeah, I assumed that the people who actually simulate (Adria/Pablo) were checking the forums. Apparently not. We should probably discuss this with them.