| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: Microbiome Immunity Project
Thread: Lots of MIP1 WUs error out |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 70
|
|
| Author |
|
|
julemand101
Cruncher Denmark Joined: Feb 28, 2020 Post Count: 7 Status: Offline Project Badges: 
|
That was also a theory I had but if you check the executable you can see it is statically linked:
[root@beta www.worldcommunitygrid.org]# file wcgrid_mip1_rosetta_7.16_x86_64-pc-linux-gnu And does therefore not have any external dependencies to any libraries (not even glibc): [root@beta www.worldcommunitygrid.org]# ldd wcgrid_mip1_rosetta_7.16_x86_64-pc-linux-gnu The issue could then be the Linux kernel where I run 5.8.12. I am going to test with 5.4.68 which are the LTS kernel version available for my OS (Arch Linux). |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7660 Status: Offline Project Badges: 
|
OK, I have completed the first 4 MIP units on a 2600K running Linux 4.4.0-21 generic x86 _64 kernel. Linux Mint 18 Sarah. all completed between 1.2 and 1.6 hours and all are valid. The other 4 jobs running were SCC. I will up the ante to 5 concurrent jobs and see hw that goes.
----------------------------------------You may be right, it may be your updated kernel. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
julemand101
Cruncher Denmark Joined: Feb 28, 2020 Post Count: 7 Status: Offline Project Badges:
|
After running a day with Linux kernel 5.4.68-1-lts I can conclude that MIP1 are still failing. I will try to see if I can get access to an even older kernel but that is not something there is official support for in Arch Linux.
----------------------------------------But I will not leave out the possibility that it is glibc since the stacktrace does contain a reference to that library (and libnss_files-2.32.so). So it could be that it is still using glibc even if it is statically linked. Oct 01 00:07:13 beta.archlinux.dk systemd-coredump[16147]: Process 14664 (wcgrid_mip1_ros) of user 969 dumped core. [Edit 1 times, last edit by julemand101 at Oct 1, 2020 8:26:30 AM] |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7660 Status: Offline Project Badges:
|
I have now run through about 20 MIP units with nary a problem. I see you are partial to the Arch distro for your use, but it may pay to look at one of the many other distros if you can not find the kernel you are seeking. It would be interesting to see if any other Arch users are experiencing this difficulty. If no other users of your distro are experiencing the same problems, it might be hardware.
----------------------------------------Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
TeamAdmin
Cruncher Joined: Aug 21, 2010 Post Count: 1 Status: Offline |
Arch user here (Endeavour OS) with AMD Ryzen 1700.
All mip1 hare failing. I also noticed reference to /usr/lib/libc-2.32.so in the dumps. |
||
|
|
mdxi
Advanced Cruncher Joined: Dec 6, 2017 Post Count: 109 Status: Offline Project Badges: 
|
Arch is a rolling release distro, so it's much closer to the leading edge of upstream development than a timed-release distro. But there's nothing exotic about it. We Arch users should probably be viewed as canaries because he issue currently affecting us is very likely going to become a problem for everyone else... eventually :)
----------------------------------------I also thought BOINC apps were statically compiled, so I had also assumed it couldn't be a library issue. There is a way to find that out... # ldd ./wcgrid_mip1_rosetta_7.16_x86_64-pc-linux-gnu Yeah, that's static, so any references to libs should be to the compiled-in version? Unless my understanding of how static binaries is off (which is, of course, possible). But in any case, libc 2.32-4 is the current build on Arch, so that shouldn't be a mismatch in any situation? So confusing. Edit: I notice that MIP1 has declined from around 500k validated WUs/day a month ago, to 150K WUs/day for the past two weeks. I don't think that's related to this specific issue, but it's interesting.  [Edit 1 times, last edit by mdxi at Oct 5, 2020 6:05:55 AM] |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7660 Status: Offline Project Badges:
|
Edit: I notice that MIP1 has declined from around 500k validated WUs/day a month ago, to 150K WUs/day for the past two weeks. I don't think that's related to this specific issue, but it's interesting. Probably more related to the increase in SCC.Since Linux is more efficient on the Vina projects than Windows, there may have been a rotation out of MIP1 to SCC by a significant number of Linux users. The Arch distro problems with MIP1 may also be a contributing factor. Update: I ran about 50 MIP units on Linux Mint in combination with other projects (SCC,ARP and OPN) and they all came back valid. Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
buscher
Cruncher Joined: Oct 3, 2011 Post Count: 5 Status: Offline Project Badges: 
|
I also have problems with MIP1, I did opt-out for a month or so, hoping that it would fix itself magically, but just tried again on 2 different PCs, but still, they all error out.
----------------------------------------Kernel: 5.8.17 / 5.9.2 Glibc: 2.32 CPU: AMD Ryzen 7 2700X / Intel i7-10610U EDIT: using Linux Gentoo btw :) [Edit 1 times, last edit by buscher at Nov 1, 2020 9:46:26 AM] |
||
|
|
Sgt.Joe
Ace Cruncher USA Joined: Jul 4, 2006 Post Count: 7660 Status: Offline Project Badges:
|
I also have problems with MIP1, I did opt-out for a month or so, hoping that it would fix itself magically, but just tried again on 2 different PCs, but still, they all error out. Kernel: 5.8.17 / 5.9.2 Glibc: 2.32 CPU: AMD Ryzen 7 2700X / Intel i7-10610U EDIT: using Linux Gentoo btw :) Have you tried a different Linux distro ? Cheers
Sgt. Joe
*Minnesota Crunchers* |
||
|
|
buscher
Cruncher Joined: Oct 3, 2011 Post Count: 5 Status: Offline Project Badges:
|
No, I have not and frankly I will not, as all other projects are working fine and I have no other problems beside this one. (All other: the rest of WCG, Rosetta, GPUGRID, ...)
|
||
|
|
|