SETI problem

[Intelligence3]

Might have been me, Smoke. I was having weird issues with one client. I will stop it for the night and reinstall tomorrow. Been a while since I installed, hope I can rmemeber how I do it.

 

[Robor]

Well I followed the advice in this thread (delete the result.sah and replace the user.sah with a valid one) and it appears as though my 12 stalled clients have fired back up. Hopefully things will are okay at work now. Will repeat the process at home after work tonight. After that my numbers should be back to normal.

Thanks for the help guys!

 

[Freewolf]

I've had two of the bad work units show up at home so far. Deleted them and evrything fine after that

 

[Intelligence3]

Anyone have the i386-winnt-cmdline version 3.03 installer? I can't find mine.

 

[Smoke]

Any know who this guy is?

AK_Yanqui

I'm about to "Torch" him.

1. I can't contact him.
2. He is not getting any WUs PASSED-THROUGH.
3. He may not be the only culprit but I don't know what else to do.

Any suggestions or advice?

Oh, yes ... I've got over a hundred WUs piled up and climbing.

 

[compudog]

Originally posted by: Intelligence3
Anyone have the i386-winnt-cmdline version 3.03 installer? I can't find mine.

I have it. PM me for email.

 

[Smoke]

Originally posted by: Intelligence3
Anyone have the i386-winnt-cmdline version 3.03 installer? I can't find mine.

It's in the link in my sig-line.

 

[lobadobadingdong]

mine weren't from bad WU's, my problem was a corrupted hard drive, my virtual memory file had a few errors, got it fixed just a few minutes ago.

 

[Spacehead]

For those with the "100" error, here's a more detailed explaination.
Link

 

[Smoke]

Thanks for that link, Spacehead.

"The TeAm SetiQueue" has been really crippled by this problem.

My Q has 112 CLIENTS, many of which are using multiple computers. It is an almost impossible task to get each of those CLIENTS to check the status of each of their machines. Many have "Set It and Forget It" ala RONCO.

I have tried changing all sorts of settings to stop this error. Here is an example taken from my LOG:

6:39am: s@h_wrk Returning passing through request response
6:39am: s@h_wrk Passthrough: Seti@home status: ErrorCode 0x00000064 100
6:39am: s@h_wrk Passing through request send_result_get_user_stats

My Q is getting many, many timeouts such as the following:

6:42am: s@h_wrk Sending Result: SendRequest to shserver2.ssl.berkeley.edu failed: Connection Timeout

I really don't know what to do. The Q is chugging away, doing its best and every once in a while a few completed WUs get sent in BUT the Queued Results keep piling up. I noticed with the reduced flow overnight, the number of queued results has shrunk but it will probably increase during the day.

I've tried rebooting the Q but that does not seem to help the situation.

I've DISABLED the "Override for pass-through operations" which does not help because these "defective WUs" are not coming from NEW CLIENTS but from OLD ESTABLISHED CLIENTS.

Last night, I noticed my primary workstation, the one I'm using to write this message, had a YELLOW SETIDRIVER ICON. Rebooting, stopping and restarting SETIDRIVER had no effect. Upon further inspection I discovered that my main workstation had one of the "defective WUs".

Now the fix is not that hard but there are many of my Q's CLIENTS that would be definitely put back by this FIX and as I already mentioned, contacting them in the first place would be very difficult for many and impossible for some.

I am monitoring the members of my mini-team and will contact those that seem to be having a problem. But even some of them are going to be beyond contact.

As Matt Lebofsky mentioned in the link provided by Spacehead, the current effort is to get BOINC going and not waste time trying to fix ongoing issues with S@H-1. So it appears my Q, The TeAm SetiQueue, will have to limp to the finish line.

As per usual, I'm all ears to any suggestions.

 

[Assimilator1]

Sorry to hear your having trouble Greg ,I see you have 18 results backed up atm ,is that a reduction?

 

[Smoke]

Yes, 18 is actually great. I've been having over 200+ backed up at times.

You might know, I found another "defective WU" on one of my own machines - Smoke #07 (my wife's computer whose desk is right across from mine). So based on my own experience, this problem has got to be wide-spread.

 

[Assimilator1]

Yeah many people have been hit by it on various teams

Glad to hear your Q has releaved itself!

 

[ICXRa]

Bump for folks that still may be finding out they got this problem.

My home fleet has been crunching right along and tonight I find a bad one!

 

[Assimilator1]

Damn their still about?