Archived

This topic is now archived and is closed to further replies.

yipikaye

segfault libc-2.11.2

Recommended Posts

Here's the final output of strace when it crashed.

I forgot to suggest that people run strace() - thanks. Unfortunately, I don't see anything about the output that helps identify the source of the crash.

I had prepared a release and had finished testing it this afternoon. However, there have been a couple of fixes today to crash-related bugs in new generic code (added since the last release). So, I'm going to take tonight's build (if unit/system testing passes) and stress test that starting tomorrow. I didn't want to trade a possible solution to one crash for two additional types of crashes.

Share this post


Link to post
Share on other sites

mcdonald, are you still unable to reproduce this error yourself? I assume this would help in diagnosing the problem :P. I can give you access to one of my VPS' that the segfault happens on. Contact me via FluxBB form email if you want to try it.

Share this post


Link to post
Share on other sites

Is there any news on this bug ?? A old version that does not crash ?? I spend more time re-hashing torrents then I do downloading them. I'll hook you up with access too if you need it ... to reproduce the crash. Sometimes it works for 24+ hours ... sometimes it dies after just hours ... never minutes tho.

Share this post


Link to post
Share on other sites

Interestingly, I have recently bought a new machine and reinstalled Gentoo from scratch and have not had it crash in the 3 days it's been running. It's exactly the same settings because I just gzipped the directory and copied it over to the new server. The only differences (besides the obvious hardware upgrade) is that I'm storing to an eSATA mounted external drive and I'm now using glibc-2.12.2 / gcc-4.5.2.

Share this post


Link to post
Share on other sites
mcdonald, are you still unable to reproduce this error yourself? I assume this would help in diagnosing the problem :P. I can give you access to one of my VPS' that the segfault happens on. Contact me via FluxBB form email if you want to try it.

Yes, I'm still not able to reproduce the error, and, yes, that would help. However, I'll wait on your offer of access to your VPS.

The release archives are built, unit tested, and system tested on Ubuntu 9.10 (VM), and stress tested on Ubuntu 10.10 server (native). The system tests probably don't run long enough to encounter the problem posters to this thread are seeing. I'm preparing an Ubuntu 10.10 VM for building future releases.

My guess is there's some sort of version incompatibility in some users' glibc's, since there appear to be a number of users for whom this problem does not occur. As a result, needing to spend my time on tasks that are likely to have the greatest benefit to the most users, I don't think I can get as far looking at other people's VMs as I can by rebranching and updating the VMs I use for building and testing. If, after all that, posters to this thread are still seeing the problem, I can dig into it further. I've sent another release to Firon; looks like he hasn't posted it yet.

Share this post


Link to post
Share on other sites
I've sent another release to Firon; looks like he hasn't posted it yet.

Firon posted it; try it out. One post in another thread said it resolved their seg fault problem.

Share this post


Link to post
Share on other sites

Ok, im trying out the new version of utorrent. Hopefully this one works much better. I think within the last 2 weeks ive spent more time hashing my torrents then downloading. Thanks guys for all the help and ideas, and for supporting a linux version of utorrent. Is there any kind of debug modes you want me to run ? strace ? time ? internal debugging options ?? if the error keeps alive.

Im running a pretty stock Debian Testing distro except for a new kernel which I compiled myself (using the stock debians .menuconfig)

Thanks again :)

chris

Share this post


Link to post
Share on other sites
Update report on build 24118: utserver uptime 1d 12h 11m so far, and no segfaults! I'll keep an eye and report any change...

Agreed! My utserver uptime (buidl 24118) is 1d 19h 46m with no segfaults. It's looking good so far..

Share this post


Link to post
Share on other sites
Update report on build 24118: utserver uptime 1d 12h 11m so far, and no segfaults! I'll keep an eye and report any change...
New version has fixed the segfaults :)

I would tend to agree, now my utserver uptime is 3d 16h 32m!!! (Still going 4d 9h 30m later...)

Thanks!!!

:cool:

Share this post


Link to post
Share on other sites

I now consider this segfault issue resolved.. 8 days uptime so far!

Thanks to the developers and everyone involved in making this great product and fixing the segfault problem!

Keep up the great work!! :D

Share this post


Link to post
Share on other sites

Awesome ... im at over 4 days uptime with it and no problems. My machine is at 25 days so a reboot was not one of the variables in it not working. Thanks alot guys for the awsome support.

Chris :D

Share this post


Link to post
Share on other sites

Not sure if it's appropriate to respond in this thread vs. creating a new one at this point, but I've had the same issue for some time and it's not been resolved, even after upgrading - most recently to 3.0-24733. I haven't been able to figure out how to reproduce it on cue, but it always seems to die after maybe a half hour or more of normal function.

/var/log/messages:


Mar 13 15:21:07 hostname kernel: [922234.272558] utserver[2424]: segfault at b32fd000 ip 080be838 sp bfa20a60 error 4 in utserver[8048000+b6000]
Mar 13 15:26:20 hostname kernel: [922547.638555] utserver[2563]: segfault at b50fd000 ip 080be838 sp bf8343e0 error 4 in utserver[8048000+b6000]
Mar 13 18:21:00 hostname kernel: [933027.176144] utserver[2595]: segfault at b43fd000 ip 080be838 sp bfeb8b80 error 4 in utserver[8048000+b6000]
Mar 13 19:01:43 hostname kernel: [935470.527619] utserver[3691]: segfault at b47ff000 ip 080be838 sp bfc563a0 error 4 in utserver[8048000+b6000]
Mar 13 20:47:03 hostname kernel: [941790.732071] utserver[3929]: segfault at b44fd000 ip 080bf2c0 sp bf9338c0 error 4 in utserver[8048000+b9000]

uname -a:


Linux hostname 2.6.32-27-generic-pae #49-Ubuntu SMP Thu Dec 2 00:07:52 UTC 2010 i686 GNU/Linux

That's all that's seemed relevant - I'll provide more if requested.

Share this post


Link to post
Share on other sites
Not sure if it's appropriate to respond in this thread vs. creating a new one at this point, but I've had the same issue for some time and it's not been resolved, even after upgrading - most recently to 3.0-24733. I haven't been able to figure out how to reproduce it on cue, but it always seems to die after maybe a half hour or more of normal function.

Oddly, after having no crashes for a few weeks of using build 24118, my utserver started crashing again in the same way you are experiencing (for no apparent reason). This makes me wonder about the cause or reason for the crash. However, after upgrading to build 24722 my utserver has not crashed at all since. I hope this is also not a short lived success.

Verify you have upgraded to the new build correctly (try a new/blank install) and maybe do some testing with various situations (seeding, downloading, not seeding or downloading, etc).

Also, as Superman710e recommended, I would suggest using monit to restart the process when it dies. It's not a good solution but it certainly helps! :D

Share this post


Link to post
Share on other sites

Oddly, after having no crashes for a few weeks of using build 24118, my utserver started crashing again in the same way you are experiencing (for no apparent reason). This makes me wonder about the cause or reason for the crash. However, after upgrading to build 24722 my utserver has not crashed at all since. I hope this is also not a short lived success.

I was still crashing on 24118 (I think the problem might have been introduced in 23090; can't remember.) Like you said, I'm wondering what the cause is - perhaps related to an outside package version used by utserver? I wouldn't know what to start comparing if so.

Verify you have upgraded to the new build correctly (try a new/blank install) and maybe do some testing with various situations (seeding, downloading, not seeding or downloading, etc).

I upgraded by simply copying the new binary over the old one (along with webui.zip; not sure if that's necessary if using built-in server.) The configs were left in place - maybe I'll try wiping those out too. It would almost be easier if the failure was immediate - as it stands I'll only be able to try a few different configurations/scenarios in a given day. :rolleyes:

Also, as Superman710e recommended, I would suggest using monit to restart the process when it dies. It's not a good solution but it certainly helps! :D

Sounds like the best solution for now, but I'm hoping I'm not alone with this so we can work toward getting the real problem fixed. :D

Share this post


Link to post
Share on other sites