OFFICIAL "SERVER PROBLEMS" THREAD

[rebirther]

Originally Posted by Madpoo

EDIT: Just FYI, I started taking a closer look at the server performance lately because we have certain users doing a LOT (and I mean a LOT!!!) of TF work, and we got a dump the other day of nearly 300,000 results at once. There's a special bulk processor in place just for this and it took several hours to process it all and we had some other jobs acting a little funny in the meantime, but I think we got it all sorted out. It pointed out some areas for improvement and got the gray matter sparking.

Cant find any reported results for SRBase in the list (except factors), can you pls check, reported around 30k+ the last 2 hours.

[Madpoo]

Originally Posted by rebirther

Cant find any reported results for SRBase in the list (except factors), can you pls check, reported around 30k+ the last 2 hours.

I see 42,531 results in the queue that were uploaded between 16:27 and 16:37 UTC.

They're in there but I need to figure out how to jumpstart that processing task.

[Madpoo]

Originally Posted by Madpoo

I see 42,531 results in the queue that were uploaded between 16:27 and 16:37 UTC.

They're in there but I need to figure out how to jumpstart that processing task.

They're processing now. I don't know if something triggered it automatically or if James or George did something, but either way, the results should start appearing.

[James Heinrich]

Originally Posted by PhilF

Possibly unrelated, but I submitted a PRP manual test result during the problem period. It was accepted and it shows the proof file was uploaded, but it seems stuck there. It hasn't been assigned for a certification check, which is something I haven't seen before. I thought certification tests were assigned almost immediately after the proof is uploaded?

It is a known issue, George and Aaron are looking into it.

[Rubiksmath]

I was reading through this thread and saw something that reminded me of an interesting situation that occurred perhaps a month ago. It was under very strange circumstances as I have a lets say undesirable situation when it comes to internet.

So what happened was I ran 2 ECM curves on M206923 (dont remember why), and basically what happened was the automatic results being sent by Prime95 were being accepted by the server, however the app seemed to think there was some error, and as such resent the data every 70 minutes, each time getting accepted, and this went unnoticed for a bit before I noticed and figured out I could stop it by deleting prime.spl, but as a result I ended up getting credit for 8 ECM curves when I only ran 2.

This hasn't happened since, I cannot quite remember the strange circumstances that probably caused this, but I do remember that they were strange, I do remember that I had done just something like submitting these results manually due to my firewall issue (probably not expecting auto results to work).

Here's the "error" code if it helps but I doubt this will happen again anytime soon given how specific the circumstances I was in were:

[Fri Oct 14 16:30:26 2022 - ver 30.8]
Sending result to server: UID: Rubiksmath/badlaptop, M206923 completed 2 ECM curves, B1=1000000, average B2=149000000, Wi4: 0D87D990, AID: D9AAB6E1DBA4E8B2CB1E5B6368BE1282

PnErrorResult value missing.  Full response was:
certpnErrorResult=0
pnErrorDetail=CPU credit is 0.2270 GHz-days.
==END==

[congsz]

Originally Posted by Rubiksmath

PnErrorResult value missing. Full response was:
certpnErrorResult=0
pnErrorDetail=CPU credit is 0.2270 GHz-days.
==END==

This issue has been reported in this thread half a month ago, you can check it out from post #2535.

[Rubiksmath]

Originally Posted by congsz

This issue has been reported in this thread half a month ago, you can check it out from post #2535.

That is slightly different in that I believe their results were not accepted (or if they were it is redundant since P-1 doesnt need to be ran multiple times with the same bounds), but in this case my results were accepted each time.

[Prime95]

Originally Posted by Rubiksmath

That is slightly different in that I believe their results were not accepted (or if they were it is redundant since P-1 doesnt need to be ran multiple times with the same bounds), but in this case my results were accepted each time.

True, but the extra reported curves won't affect future ECM assignments very much.

[S485122]

Originally Posted by S485122

2022-11-01 23:21 is the last time the "GIMPS Assignment Rules" were updated.

The assignment expirations scripts are also stalled.

[rebirther]

The upload script has stopped working for SRBase, can you pls check?

[James Heinrich]

Originally Posted by rebirther

The upload script has stopped working for SRBase, can you pls check?

It was paused briefly while doing a server update, as far as I know it should be working again. If not, please email me directly with details of what's "not working".