Possible Splitter Problem?

Message boards : Number crunching : Possible Splitter Problem?
Message board moderation

To post messages, you must log in.

AuthorMessage
Bob Giel
Volunteer tester

Send message
Joined: 11 Jan 04
Posts: 76
Credit: 5,419,128
RAC: 0
United States
Message 1092059 - Posted: 31 Mar 2011, 21:30:03 UTC

For the last few weeks, I've been getting the following indicated error when the client requests additional work units. When you look at the workunit detail in your account it states "can't find input file". Are the splitters failing to write date to the file?

SETI@home 3/31/2011 4:19:49 PM [error] MD5 check failed for 20no10ab.6543.5793.10.10.253
SETI@home 3/31/2011 4:19:49 PM [error] expected 13c864fc0b047a11506d2b236e0e0ec9, got f8a985a6df916e4e44573b434fd562d6
SETI@home 3/31/2011 4:19:49 PM [error] Checksum or signature error for 20no10ab.6543.5793.10.10.253


ID: 1092059 · Report as offensive
bill

Send message
Joined: 16 Jun 99
Posts: 861
Credit: 29,352,955
RAC: 0
United States
Message 1092077 - Posted: 31 Mar 2011, 23:03:16 UTC - in response to Message 1092059.  

Is this happening on all your machines?
ID: 1092077 · Report as offensive
Bob Giel
Volunteer tester

Send message
Joined: 11 Jan 04
Posts: 76
Credit: 5,419,128
RAC: 0
United States
Message 1092125 - Posted: 1 Apr 2011, 3:20:12 UTC - in response to Message 1092077.  

Yes, and only when the machines request multiple tasks.
ID: 1092125 · Report as offensive
Profile Donald L. Johnson
Avatar

Send message
Joined: 5 Aug 02
Posts: 8240
Credit: 14,654,533
RAC: 20
United States
Message 1092175 - Posted: 1 Apr 2011, 7:04:24 UTC - in response to Message 1092059.  

For the last few weeks, I've been getting the following indicated error when the client requests additional work units. When you look at the workunit detail in your account it states "can't find input file". Are the splitters failing to write date to the file?

SETI@home 3/31/2011 4:19:49 PM [error] MD5 check failed for 20no10ab.6543.5793.10.10.253
SETI@home 3/31/2011 4:19:49 PM [error] expected 13c864fc0b047a11506d2b236e0e0ec9, got f8a985a6df916e4e44573b434fd562d6
SETI@home 3/31/2011 4:19:49 PM [error] Checksum or signature error for 20no10ab.6543.5793.10.10.253

The few times I've seen errors like those, it was actual file corruption during download. I had a bad connection in the phone junction box, it was messing up "everything". Had to replace 20 feet of 50-year-old phone wire.
Donald
Infernal Optimist / Submariner, retired
ID: 1092175 · Report as offensive
bill

Send message
Joined: 16 Jun 99
Posts: 861
Credit: 29,352,955
RAC: 0
United States
Message 1092187 - Posted: 1 Apr 2011, 8:43:06 UTC - in response to Message 1092125.  

Are you running off of a switch or a hub?
ID: 1092187 · Report as offensive
Bob Giel
Volunteer tester

Send message
Joined: 11 Jan 04
Posts: 76
Credit: 5,419,128
RAC: 0
United States
Message 1092227 - Posted: 1 Apr 2011, 16:57:53 UTC - in response to Message 1092187.  
Last modified: 1 Apr 2011, 16:58:11 UTC

Are you running off of a switch or a hub?

All machines are connected to a switch which in turn connects to a cable modem. Have already upgraded the switch firmware. Cable modem firmware is up to date.
ID: 1092227 · Report as offensive
Josef W. Segur
Volunteer developer
Volunteer tester

Send message
Joined: 30 Oct 99
Posts: 4504
Credit: 1,414,761
RAC: 0
United States
Message 1092278 - Posted: 1 Apr 2011, 19:43:08 UTC - in response to Message 1092059.  

For the last few weeks, I've been getting the following indicated error when the client requests additional work units. When you look at the workunit detail in your account it states "can't find input file". Are the splitters failing to write date to the file?

SETI@home 3/31/2011 4:19:49 PM [error] MD5 check failed for 20no10ab.6543.5793.10.10.253
SETI@home 3/31/2011 4:19:49 PM [error] expected 13c864fc0b047a11506d2b236e0e0ec9, got f8a985a6df916e4e44573b434fd562d6
SETI@home 3/31/2011 4:19:49 PM [error] Checksum or signature error for 20no10ab.6543.5793.10.10.253

To answer the subject question directly, it's not a splitter problem when other hosts can download the WU and crunch it successfully.

Unfortunately, any file which fails the MD5 check is deleted very soon, so it's difficult to check what kind of corruption has happened. The "only when the machines request multiple tasks" aspect is curious. Perhaps it would be worthwhile trying a cc_config.xml file with <max_file_xfers_per_project>1</max_file_xfers_per_project> set as a way of determining if something in the download path is mixing parts of one WU with another.

You might also try <http_1_0>1</http_1_0> just in case your ISP has a proxy which needs that.
                                                                  Joe
ID: 1092278 · Report as offensive
bill

Send message
Joined: 16 Jun 99
Posts: 861
Credit: 29,352,955
RAC: 0
United States
Message 1092305 - Posted: 1 Apr 2011, 21:39:41 UTC - in response to Message 1092227.  

I'm thinking it's a networking problem. Have/Can you try taking out the switch and other PCs and just hooking up the machine with the worst problem directly to the cable modem and see if you can duplicate the problem. if not repeat the test with the other machines individually.
ID: 1092305 · Report as offensive
Bob Giel
Volunteer tester

Send message
Joined: 11 Jan 04
Posts: 76
Credit: 5,419,128
RAC: 0
United States
Message 1092360 - Posted: 2 Apr 2011, 1:37:01 UTC - in response to Message 1092278.  

For the last few weeks, I've been getting the following indicated error when the client requests additional work units. When you look at the workunit detail in your account it states "can't find input file". Are the splitters failing to write date to the file?

SETI@home 3/31/2011 4:19:49 PM [error] MD5 check failed for 20no10ab.6543.5793.10.10.253
SETI@home 3/31/2011 4:19:49 PM [error] expected 13c864fc0b047a11506d2b236e0e0ec9, got f8a985a6df916e4e44573b434fd562d6
SETI@home 3/31/2011 4:19:49 PM [error] Checksum or signature error for 20no10ab.6543.5793.10.10.253

To answer the subject question directly, it's not a splitter problem when other hosts can download the WU and crunch it successfully.

Unfortunately, any file which fails the MD5 check is deleted very soon, so it's difficult to check what kind of corruption has happened. The "only when the machines request multiple tasks" aspect is curious. Perhaps it would be worthwhile trying a cc_config.xml file with <max_file_xfers_per_project>1</max_file_xfers_per_project> set as a way of determining if something in the download path is mixing parts of one WU with another.

You might also try <http_1_0>1</http_1_0> just in case your ISP has a proxy which needs that.
                                                                  Joe

Thanks for the info, Joe, will try the cc_config.xml and see what happens.
ID: 1092360 · Report as offensive
Bob Giel
Volunteer tester

Send message
Joined: 11 Jan 04
Posts: 76
Credit: 5,419,128
RAC: 0
United States
Message 1092364 - Posted: 2 Apr 2011, 1:54:47 UTC - in response to Message 1092305.  

I'm thinking it's a networking problem. Have/Can you try taking out the switch and other PCs and just hooking up the machine with the worst problem directly to the cable modem and see if you can duplicate the problem. if not repeat the test with the other machines individually.


I download updates to utilities I use that average in the megabytes and never have a problem. Seti work files are 367 kilobytes average. I have MediaCom as an ISP and know that since they started offering VoIP, they run out of IP addresses and lock up their servers.

Thanks for the advice. Thought someone else might be having the same problem but I guess not.
ID: 1092364 · Report as offensive

Message boards : Number crunching : Possible Splitter Problem?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.