SETI problem

[Doomer]

This morning I found 2 of my herd not running SETI. On both, when I started SETI, it would say sending results for a couple of seconds then it would close without any error message or nothing. They both are doing this even after a reboot.

Anyone know what might be going on ?

Thanks

 

[Hellburner]

I think Berkeley's off the net...

 

[Doomer]

Hmmm, ok. I usually get "cannot connect to server" when this happens.

Thanks.

 

[Freewolf]

try using a q

 

[Freewolf]

If you're using setidriver check the number of work units it is set to cache. If that number is smaller then the number it currently has it will send the result back but will not download a new result to process.

 

[Slaughter]

Same problem here. Won't even connect to my queue. It is a sad day.

EDIT:

So I figured I should look at my queue.

9:27am: s@h_wrk Returning passing through request response
9:27am: s@h_wrk Passthrough: Seti@home status: ErrorCode 0x00000064 100
9:27am: s@h_wrk Passing through request send_result_get_user_stats

 

[Freewolf]

No know what to tell you. Just checked our q and everything is fine with it. Doesn't even have any work units waiting to be uploaded

 

[Doomer]

Just using the bare CLI client and SETISpy. Still doing it.

I've got another box that's fixing to dump. I'll watch it and see what happens.


....... The other box sucessfully dumped and DL'ed another WU so I must have gotten a couple of bad WU.

So now how do I dump them ? What files do I need to delete ?

Thanks

 

[Smoke]

IMHO, It would probably be best to just trash everything and start over with a fresh installation. You'll end up wasting more time trying to take half-way measures than it would be worth.

 

[Doomer]

OK, I deleted about 3 files and it then DL'ed a new WU and seems to be functioning normally again.

Next time this happens I'll know it's not a server problem and won't waste time waiting for a resolution at the server end.

Thanks all.

 

[Robor]

Hmmm... Something is wrong here. I've got 2 SetiQ's with clients behind them. One at home and one here at work. On both SetiQ's I've got clients that haven't completed a WU in the normal time and they *do not* have a WU pending. All of the clients are running SETI as a service. I restarted the SetiQ and tried to stop/start the service and nothing. This sucks. I don't have the time (or desire) to go around un/reinstalling the service. I hope it fixes itself, otherwise my production is going to drop like a rock.

Just checked... I've got 12 clients that do on have WU's at the moment. That's nearly 25% of my fleet. Plenty of WU's in the queue and none waiting to send. Here's the error message in seti_service.log...

"10/27/03 3:21:00 WARNING: SETI client died with exit code 100. Attempting to restart.
10/27/03 3:21:01 INFO: SETI client successfully started.
10/27/03 3:25:24 WARNING: SETI client died with exit code 100. Attempting to restart.
10/27/03 3:25:25 INFO: SETI client successfully started.
10/27/03 3:25:29 WARNING: SETI client died with exit code 100. Attempting to restart.
10/27/03 3:25:30 INFO: SETI client successfully started.

cut out 280K of the same error code over and over

10/27/03 13:13:31 WARNING: SETI client died with exit code 100. Attempting to restart.
10/27/03 13:13:32 INFO: SETI client successfully started.
10/27/03 13:14:14 WARNING: SETI client died with exit code 100. Attempting to restart.
10/27/03 13:14:15 INFO: SETI client successfully started.
10/27/03 13:14:45 WARNING: SETI client died with exit code 100. Attempting to restart.
10/27/03 13:14:46 INFO: SETI client successfully started."

 

[OhioDude]

I'm seeing the same thing here. Just started this morning. I have two systems that are behaving the same way. Both are running as a service. I just went to one of them, deleted everything out of the seti folder except user_info.sh, ran the client manually and got it to download a new wu from my internal Seti queue. It started processing, but when I look at my setiqueue, it shows that that machine has no wu's pending...

 

[OhioDude]

Just checked and now I've got another system hung.

The status of Berkeley should have nothing to do with my clients getting wu's. I have 14 clients that get their wu's from my internal setiqueue. I have approximately 381 wu's waiting to be distributed to my clients.

This doesn't make any sense...

Bad batch of wu's, maybe?

 

[Smoke]

This all may have something to do with S@H ... didn't we have an outage earlier today? HB mentioned this earlier in this thread.

 

[Wiz]

I had one of these last week, I use a combo of setiQ here on my main machine with Seti Driver on all the rest so each has it's own cache plus the cache of the Q.
One of the machines was hung, setidriver had a WU to send in and I was getting that error over and over in setiQ.
I ended up deleting that WU that was stuck and then everything went normally afterwards.

I guess I'm lucky I only had one! It would be nicer if this was detected before the WU processes completely, but I know we are not going to get any more development on the current set of seti tools

 

[Doomer]

Looks like I screwed something up. All my user info is gone. I opened user.sah and there is nothing in it. Lookes like I should have done what Smokeball suggested.

 

[IsOs]

My SETIDriver have been stalling since early this month. Usually, 1 or 2 machine is normal, but sometime like 1 week ago, most of my SETIDriver stalled. I deleted and reinstall. Just this weekend, 2 machines stalled again.

I think I've lost some completed workunits as well. My SETIQ supposedly submitted more than what I saw credited in my SETI Account.:frown:

Summitting workunits took longer than usual as well.

 

[Assimilator1]

I saw people at TPR forums say the same thing ,they latter found out that the results did eventually get through.

No idea about the SETi 'stalling ' problem!

 

[Smoke]

FATWU

 

[IsOs]

Originally posted by: Smokeball
FATWU

You mean they were too big to go through the needle hole!

I better have them go through their daily exercise twice

 

[Smoke]

Remember Hawaii Five-0?

They had a character named "WO FAT". lol

He was fat too. ROFLMAO

 

[Orange Kid]

Same problems here....

The finished result is to big and will not be accepted....

Just delete the result.sah and re-start the client....

Yes, it was a waste of time, but these things happen....

Bad WU's from Berkeley most likely culprit

 

[Orange Kid]

Problems continue..............

Wasting too much time..............

Going to something else.....................

Bad WU's bad bad bad WU's :brokenheart::disgust::|

 

[IsOs]

Is it possible that these workunits were meant for the new BONIC SETI version?

 

[Spacehead]

I think BOINC gets it's WU's from a different server.


From the S@H front page

? October 27, 2003 ?
If you are unable to connect due to "100" errors (or perhaps "41") you may need to exit SETI@home, remove the files "result.sah" and "user_info.sah" from your system, and then restart SETI@home. This is probably fallout from the user database crash a few days ago. Sorry for the inconvenience.