Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
World Community Grid Forums
Category: Completed Research Forum: Smash Childhood Cancer Thread: Some errors |
No member browsing this thread |
Thread Status: Active Total posts in this thread: 18
|
Author |
|
Macroman
Advanced Cruncher Joined: Jun 4, 2005 Post Count: 112 Status: Offline Project Badges: |
I have had 2 errors that appear to fit this pattern with 30 completed. Not a great average but I'm guessing we're going thru an adjustment process.
|
||
|
Selmiina
Cruncher Joined: Jan 30, 2014 Post Count: 2 Status: Offline Project Badges: |
Hey,
----------------------------------------I'm having errors with multiple machines with the following error. They are not overclocked etc, OS is Clear Linux. Other projects are running without issues. Any ideas what's wrong? Result Name: SCC1_ 0003702_ Prdm-A_ 78718_ 0-- <core_client_version>7.14.2</core_client_version> <![CDATA[ <message> process got signal 11</message> <stderr_txt> SIGSEGV: segmentation violation </stderr_txt> ]]> [Edit 1 times, last edit by Selmiina at Mar 13, 2020 11:32:35 PM] |
||
|
Macroman
Advanced Cruncher Joined: Jun 4, 2005 Post Count: 112 Status: Offline Project Badges: |
My error results were both from batch 3702 issued on 3/9 and 3/10. I now have about 110 results in the results status page with no errors since then.
|
||
|
Pandelta
Advanced Cruncher Joined: Jun 24, 2012 Post Count: 55 Status: Offline Project Badges: |
I have the exact same issue as you Selmiina. I can run ARP and MIP without issue but cannot run SCC and TB. SCC almost instantly crashes with an error and has the same text as you, given below. There must be something about those projects that don't like Intel's optimized OS. I am running on dual Xeon CPU E5-2673 v4 @ 2.30GHz.
Clear Linux OS [5.4.6-883.native|libc 2.30 (Clear Linux Software for Intel Architecture)] Result Name: SCC1_ 0003765_ Prdm-B_ 45692_ 0-- <core_client_version>7.14.2</core_client_version> <![CDATA[ <message> process got signal 11</message> <stderr_txt> SIGSEGV: segmentation violation </stderr_txt> ]]> |
||
|
Selmiina
Cruncher Joined: Jan 30, 2014 Post Count: 2 Status: Offline Project Badges: |
So there might be somekind of issue here? Is this something that should be asked from Intel devs or WCG/SCC devs?
----------------------------------------I'm not an expert on Linux (I just use it), so I don't know if there is something that user like me could do. [Edit 2 times, last edit by Selmiina at Mar 22, 2020 6:35:14 AM] |
||
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 885 Status: Offline Project Badges: |
So there might be somekind of issue here? Is this something that should be asked from Intel devs or WCG/SCC devs? I'm not an expert on Linux (I just use it), so I don't know if there is something that user like me could do. I think this is most likely to be something to do with modern Linux kernels disabling the old vsyscall mechanism; it was a mechanism for making a few extremely common system calls (such as current date/time) possible without full system call overheads, but it mapped data in fixed memory locations so was considered a possible attack vector - another mechanism to achieve the same sort of performance but with better security was implemented a long time ago now, but its use was not made mandatory at the time as it would've immediately broken a lot of programs! Over the last couple of years most of the older applications here have been rebuilt with newer versions of the standard C libraries which know the newer ways of doing system access that used to use vsyscall, but it appears that SCC1 has not been rebuilt (perhaps because SCC was on hiatus when this problem really came to the forefront. Some Linux versions (Ubuntu, for instance) have a boot-time option that manages vsyscall availability - at present Ubuntu allows its emulation but I think Debian Buster turns it off by default. However, there may be systems where it is off and you can't get it back... For more about this with particular respect to WCG, use the site search facility to look for vsyscall on all forums. Here are a couple of examples which might be of help... Firstly, one from the Suggestions/Feedback forum , which gives some information about vsyscall. It has a reference to an updated version of SCC1 going through Beta but I think that came to naught - I notice we are currently still running v7.08, not v7.16 or later. Then there's this from the BOINC Agent Support forum which (eventually) discusses trying to re-enable vsycalls. Hope this is of some use. Cheers - Al. P.S. If Tech Support see this can someone enlighten us as to the status of a newer version of the SCC1 binaries? |
||
|
armstrdj
Former World Community Grid Tech Joined: Oct 21, 2004 Post Count: 695 Status: Offline Project Badges: |
We are working on getting a new Linux build into testing. We will try and get this into beta as soon as we can.
Thanks, armstrdj |
||
|
alanb1951
Veteran Cruncher Joined: Jan 20, 2006 Post Count: 885 Status: Offline Project Badges: |
We are working on getting a new Linux build into testing. We will try and get this into beta as soon as we can. Thanks, armstrdj Thanks for the information. I'm also guessing that a build with a newer compiler might be a bit quicker too! Cheers - Al. |
||
|
|