SETI - Crucher crashing, need guidance

[compudog]

I have a cruncher that crashes at the completion (or start of the next) WU. It is a Duron 1.4 on a Shuttle AK37GT with 256MB DDR running Win2K sp4 with Seti Driver (CLI version 3.03 i386-winnt-cmdline) I set client priority to low, normal and high with the same result. The "machine" is part of a crack rack and heat is NOT a problem. It's sole purpose is for TeAm DC projects and runs 24/7 (except when I have to reboot due to crash,) There are no screen savers, the machine is used for nothing else. Any ideas? I may switch it over to DPAD if I cannot resolve the SETI issue.


Thanks.

 

[kamper]

just out of curiosity: how many cached wu's do you have in the directory with the driver?

 

[Smoke]

There is just a little extra load placed on the computer when one WU is completed and the results uploaded and a new WU downloaded.
I have seen many different setups crash during this time phase.

It usually happens to computers that are overclocked just a tad too much. You didn't mention if you were overclocking this duron system?

 

[compudog]

No OC, there are twelve WU cached. It's not set to auto-transmit. It ran DPAD overnight with no problems. I went to upload the three results it had managed to complete and it crashed again as it tried to update the Seti Driver applet (Waiting to Process, Need to fill, Ready to Transmit.) Later today I will load a fresh copy of Seti-Driver, grab some new WU's and see if that help. I may also "defrag" the HDD, but it shouldn't need it as the machine isn't used for anything but DC projects.

 

[Smoke]

Strange that it only happens with Seti. I don't think Seti uses any more of the CPU than any other DC?

Let's us know what happens next.

 

[Hellburner]

If there are no failed/failing fans on the MB or power supply, then I check for swollen/burned out caps on the MB and power supply. I've also seen various thermal transfer system failures between the HS and CPU causing weird instabilities. HD's have also been a reliability barrier, error checking the disk should give hints of any problem there. Have you recently installed any service packs? I've experience network related service pack update hell with the older NT code base. That about covers my bag of tricks for this issue...

 

[Unforgiven]

2 things....

1) many people have issues with 2k sp4 so you would definately not be alone. try rolling back to sp3 if you have the ability to in order to see if the issue is windows itself.

2) try another version of seti driver. ive had issues on some machines with the newer versions and just reverted back to older ones and it seemed to fix my issues.

 

[CXGJarrod]

Originally posted by: Unforgiven
2 things....

1) many people have issues with 2k sp4 so you would definately not be alone. try rolling back to sp3 if you have the ability to in order to see if the issue is windows itself.

2) try another version of seti driver. ive had issues on some machines with the newer versions and just reverted back to older ones and it seemed to fix my issues.

I would agree with number 1. I switched all my seti machines back to seti hide after seti driver kept crashing on me. Have not had a problem since. (all Win2k SP4 machines)

 

[compudog]

The SETI driver problem persists. Only at the completion of a WU. If I just run the CLI without SETI driver it seems OK. I have the board running DPAD now. I looked it over well and found no bulging caps, checked voltages under load, all are within spec. NTFS volume has 74% free space with few fragments. I can't roll back to SP3 as I did not archive files, though I can just reload the OS and leave it at SP2 and see if that works. I may try this tomorrow, but will leave it crunch DPAD tonight. Some good ideas here though... keep 'em coming!