What's with this?

Message boards : Number crunching : What's with this?

To post messages, you must log in.

AuthorMessage
Profile Nite Owl
Avatar

Send message
Joined: 2 Nov 05
Posts: 87
Credit: 3,019,449
RAC: 0
Message 9503 - Posted: 21 Jan 2006, 4:23:24 UTC
Last modified: 21 Jan 2006, 4:24:52 UTC

1/20/2006 11:16:49 PM|rosetta@home|Reason: To fetch work
1/20/2006 11:16:49 PM|rosetta@home|Requesting 43200 seconds of new work
1/20/2006 11:17:24 PM|rosetta@home|Scheduler request to https://boinc.bakerlab.org/rosetta_cgi/cgi succeeded
1/20/2006 11:17:24 PM|rosetta@home|Message from server: No work sent
1/20/2006 11:17:24 PM|rosetta@home|Message from server: (reached daily quota of 2 results)
1/20/2006 11:17:24 PM|rosetta@home|No work from project


Preference = 300 (74%)

Join the Teddies@WCG
ID: 9503 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Divide Overflow

Send message
Joined: 17 Sep 05
Posts: 82
Credit: 921,382
RAC: 0
Message 9505 - Posted: 21 Jan 2006, 4:55:06 UTC
Last modified: 21 Jan 2006, 5:07:19 UTC

You have your computers hidden, so I can't be sure. It sounds like you returned a *LOT* of bad results from that machine, dropping your daily quota for that host down to just 2. The project won't let you have any more work until you return a valid result, which will double your daily quota each time. Each valid result returned from that host will double your daily quota.

What is going on with that host?!?
ID: 9505 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Nite Owl
Avatar

Send message
Joined: 2 Nov 05
Posts: 87
Credit: 3,019,449
RAC: 0
Message 9507 - Posted: 21 Jan 2006, 5:07:01 UTC - in response to Message 9505.  

You have your computers hidden, so I can't be sure. It sounds like you returned a *LOT* of bad results from that machine, dropping your daily quota for that host down to just 2. The project won't let you have any more work until you return a valid result, which will double your daily quota each time.

What is going on with that host?!?

Thanks Dave, I have had mucho grande bad results, on all 30 machines... I've driving myself nuts running around the house trying to get some viable work... About half my machines right now have work to crunch. Just how viable it is remains to be seen...
ID: 9507 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Divide Overflow

Send message
Joined: 17 Sep 05
Posts: 82
Credit: 921,382
RAC: 0
Message 9508 - Posted: 21 Jan 2006, 5:09:05 UTC

Each valid result returned from that host will double your daily quota, so as soon as you get your problems sorted out, you should be back up to full quota again in next to no time. Good luck!
ID: 9508 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Nite Owl
Avatar

Send message
Joined: 2 Nov 05
Posts: 87
Credit: 3,019,449
RAC: 0
Message 9509 - Posted: 21 Jan 2006, 5:23:21 UTC

@Dave, I made my Computers visable so you could take a look...
ID: 9509 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Divide Overflow

Send message
Joined: 17 Sep 05
Posts: 82
Credit: 921,382
RAC: 0
Message 9510 - Posted: 21 Jan 2006, 5:44:50 UTC
Last modified: 21 Jan 2006, 6:01:39 UTC

Wow! I wish I had that much computing power available...

It certainly looks like this guy is having the most problems:
https://boinc.bakerlab.org/rosetta/show_host_detail.php?hostid=53940

I see a lot of your computers have a download error for their bad WU's. Are you having connectivity problems? Did you change any proxy addressing recently?

To answer your specific question, what's going on? When a host returns a result with errors, it's daily quota is reduced. This is done to prevent unattended rogue bad systems out there from draining the total number of WU's available to everybody. As soon as a system that was having temporary problems starts returning good results again, the quota avaialable to it doubles with each good result it returns.

I can't tell what is causing you to return bad results at this point, but the download errors are a good clue to start from.
ID: 9510 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Nite Owl
Avatar

Send message
Joined: 2 Nov 05
Posts: 87
Credit: 3,019,449
RAC: 0
Message 9511 - Posted: 21 Jan 2006, 6:31:48 UTC
Last modified: 21 Jan 2006, 6:48:57 UTC

If you look at a few more machines you'll there's others that have had more bad results.... Both download and during processing.... I use to be in 11 place in the stats with over 5900 RAC. I started dropping like I was pole axed, so I upgraded my wireless router and got a new, faster two way satellite adapter and service (DW7000) with Professional software, to see whether that was the problem or not.... It wasn't... Evidently your files are too sensitive and can't handle the trip to the satellite and back!
Join the Teddies@WCG
ID: 9511 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Nite Owl
Avatar

Send message
Joined: 2 Nov 05
Posts: 87
Credit: 3,019,449
RAC: 0
Message 9518 - Posted: 21 Jan 2006, 8:56:01 UTC - in response to Message 9511.  

If you look at a few more machines you'll there's others that have had more bad results.... Both download and during processing.... I use to be in 11 place in the stats with over 5900 RAC. I started dropping like I was pole axed, so I upgraded my wireless router and got a new, faster two way satellite adapter and service (DW7000) with Professional software, to see whether that was the problem or not.... It wasn't... Evidently your files are too sensitive and can't handle the trip to the satellite and back!

Here is an example of a download error:



Notice the percentage received? 1267.47%? No wonder it fails... 99.9% of everything connected to FATSA is just about guaranteed to fail, after all other associated files have downloaded, although that particular job was replaced with another 1n0u with a FASTA file (the 00.1%) and it downloaded fine...

Join the Teddies@WCG
ID: 9518 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Snake Doctor
Avatar

Send message
Joined: 17 Sep 05
Posts: 182
Credit: 6,401,938
RAC: 0
Message 9570 - Posted: 22 Jan 2006, 1:05:48 UTC

Nite Owl,

I just stumbled on this thread, and took a quick look at some of the errors you are getting. Correct me if I am wrong, but it looks like you are running BOINC client 5.2.6. If so you may need to upgrade for R@H. A lot of us were getting similar errors running versions of the BOINC client below 5.2.8. At this point the "blessed" version is 5.2.13. 5.2.8 solved some of the problems you are seeing. but 5.2.13 is better.

In my experience 5.2.13 is a bit more stable but I am running the GUI version so I can't speak to CLI or menubar versions. With the number of machines you have you may want to just pick the one having the most trouble and try upgrading that one first. Some of the fixes in 5.2.13 were for downloading issues. Your mileage may vary.

You may already know this but you may want to drain your queue before turning off BOINC or any WUS you have may abort making your problems worse.

Let us know how this turns out for you. The errors you are seeing are similar to some I have seen before and the project folks are working on that specific problem.

Regards
Phil

We Must look for intelligent life on other planets as,
it is becoming increasingly apparent we will not find any on our own.
ID: 9570 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Profile Nite Owl
Avatar

Send message
Joined: 2 Nov 05
Posts: 87
Credit: 3,019,449
RAC: 0
Message 9580 - Posted: 22 Jan 2006, 9:20:53 UTC

I have version 5.2.13 running on most of the problem machines already... Read somewhere that if you have a problem use .13... so I did... Thanks for your reply SD... <waves>
Join the Teddies@WCG
ID: 9580 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
FluffyChicken
Avatar

Send message
Joined: 1 Nov 05
Posts: 1260
Credit: 369,635
RAC: 0
Message 9581 - Posted: 22 Jan 2006, 11:47:34 UTC

No harm in trying 5.2.15 either

Link including 5.2.15

Although I see nothing in the changelog that may help you :( (probably as I don't know what's causing the problem :D

Changes since 5.2.14
Use curl error messages for network errors. Maybe more useful information will be displayed than what was previously displayed.
Fixed another problem which could prevent BOINC Manager from connecting to core client.
Mac only: we are now building with newer tools and libraries: XCode 2.2, WxWidgets 2.6.2, libcurl 7.15.1.
Mac only: in anticipation of Intel-based Macs, the Mac BOINC Manager package name now specifies the cpu architecture: boinc_5.2.15_macOSX_powerpc.zip.
Changes since 5.2.13
The manager should now correctly detect if the core client is currently running on the machine.
Mac: Only administrators can attach to a new project (for improved security).
Mac: No longer crashes if there is a space character in the path to BOINCManager.

Team mauisun.org
ID: 9581 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
Astro
Avatar

Send message
Joined: 2 Oct 05
Posts: 987
Credit: 500,253
RAC: 0
Message 9582 - Posted: 22 Jan 2006, 13:20:16 UTC

Dr. David Anderson posted a email about the Alpha testing of 5.2.15, here part of it is.

David Anderson to boinc_alpha
More options Jan 17 (5 days ago)

Testing of 5.2.15 is complete (I think - is it, Rom?)
It didn't pass.



ID: 9582 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote
AMD_is_logical

Send message
Joined: 20 Dec 05
Posts: 299
Credit: 31,460,681
RAC: 0
Message 9584 - Posted: 22 Jan 2006, 13:50:30 UTC - in response to Message 9518.  

Here is an example of a download error:



Notice the percentage received? 1267.47%? No wonder it fails... 99.9% of everything connected to FATSA is just about guaranteed to fail, after all other associated files have downloaded, although that particular job was replaced with another 1n0u with a FASTA file (the 00.1%) and it downloaded fine...

My guess is that there is some software on your systems that has inserted itself between the boinc program and the internet. It lets .gz files through, but runs .fasta files through some filter which messes things up. Perhaps this software can be told to let .fasta files through without altering them.
ID: 9584 · Rating: 0 · rate: Rate + / Rate - Report as offensive    Reply Quote

Message boards : Number crunching : What's with this?



©2024 University of Washington
https://www.bakerlab.org