that can be a motherboard related problem as well and/or a heating issue...

Segmentation fault in Ubuntu 12.04.3 LTS
 

Hi,

I'm getting segmentation fault on my system. I'm very fed-up of this.

System Specs:

OS: Ubuntu 12.04.3 LTS (with all the updates installed)
Arch : 64 bit


I already checked RAM and HDD by doing Memtest and fsck respectively. No errors found on those devices.

Recently I opened synaptic package manager and it was immediately crashed. So dmesg says:

http://pastebin.com/SfSsKJnY

and I ran a strace synaptic from command line as root to capture strace for the crash. You can find the report here:

http://pastebin.com/UFBZJkWz

My whole system is getting crashed due to this and all the applications are affected to this issue.

I have already posted on thread , but didn't resolve the issue. For your reference, my old thread link is http://www.linuxquestions.org/questi...me-4175482101/

Please help.

Please do not open several threads on the same topic. I have merged your threads.

Thank you. I thought I can open new thread because I got new set of logs for the people to analyze. If you can , please update my last post to the first post of this thread so that people can see it.

Many times segfaults are the result of different version libraries. Do you use only the default repositories?

Yes. Only the default repo.

I even re-installed and tried another distro's and still I'm getting segmentation fault.

Initial thoughts.



Installed a 64 bit distro on a 32 bit machine?

$ cat /proc/cpuinfo | grep -i " lm "

If you don't have " lm " you don't have a 64 bit cpu. For x86 machines anyway.



Running out of RAM.

As in too many applications and stuff. If you don't have swap space things are hardwired to save the kernel first, so things get stopped and rather harshly to free up resources and keep the kernel functioning. Swap space functions as virtual ram.

$ sudo dd if=/dev/zero of=swap_file bs=1M count=1000
$ sudo mkswap -v 1 swap_file
$ sudo swapon swap_file
$ sudo swapon -s

(for 1GB of swap on a file in a file system, not the most efficient way)
(you'll have to swapon swap_file at each boot if you don't add it to /etc/fstab)

If you run out of RAM, you'll get random application crashes to free up RAM. By design. You might play with the swappiness value to tune your system with how aggressively it uses swap. The default 60 is pretty disk I/O intensive. But doing less multi-tasking and not maxing your RAM usage avoids almost all of this.



Or maybe you're out of free space on your hard drive. The /tmp location is sort of pre-emptive swap and can chew up available drive space. Along with archived debs from updates, and browser cache, and other stuff.



These may not be your issue. But your symptoms seem to mirror the expected behaviors for issues like those above.

Thanks for the reply. Here are my results:

1) I have lm as you mentioned

2) RAM is 4 GB and swap is 4GB and the system shows enough free memory.

3)I have checked the Hard drive space and it is proper. I even fsck'ed using live cd and it showed no block errors and I did a RAM check using Memtest twice and the RAM it passed twice.

Here's my Firefox's strace log: (new)

http://pastebin.com/DMGxjpXf

Please help me to resolve the issue.

The common thread is (Resource temporarily unavailable), so the question is which resource?

Are you running things off of a network (like the filesystem) and the network is quirky? If you're running off of a USB drive the usb cable or port could be having issues. Try another cable and/or port. Otherwise it could be a bad install or a faulty drive. You can --reinstall some packages and they might get replaced to a good part of the drive. And stop being symptomatic. For browsers sometimes you just have to delete or rename their .mozilla or equivalent configurations and restart them. To keep old configs from breaking the newer versions of the application.

No network file system and no USB drive. I re-installed the operating system many times and the same continues. I even tried deleting .mozilla directory. This segfault appears almost with any applications.

Are you using tmpfs mounts anywhere?

One cause of segfaults is having a malloc failure. Malloc itself no longer returns NULL when memory is tight/unavailable. It doesn't fail until you actually try to use the memory. If you feel you have sufficient memory + swap, you could try disabling oversubscription for testing.

If you get a failure, it should be due to an OOM message getting reported.

You COULD (this is very uncertain) be getting errors from swap (fsck doesn't test swap).

Seemingly random behaviors most often times indicate a hardware issue. Although if you're running ia_64 as in Intel Architecture on an AMD cpu, that might be a factor. There have been a few defective CPUs out there and motherboard firmware type issues. Have you tried running a 32 bit version? If you have and get the same issues there, it might be hardware related. If you have and don't have the same issues, it is likely software related.

Here is my df -h:

# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/sda3 314G 48G 251G 17% /
udev 1.9G 4.0K 1.9G 1% /dev
tmpfs 1.9G 1.9M 1.9G 1% /tmp
tmpfs 771M 808K 771M 1% /run
none 5.0M 8.0K 5.0M 1% /run/lock
none 1.9G 156K 1.9G 1% /run/shm
cgroup 1.9G 0 1.9G 0% /sys/fs/cgroup

Btw, my system is having 4 GB RAM and so the swap is not being used until I run very heavy applications.

In reply to shadow_7:

Previously I had 64 bit Deepin linux and 64 bit Windows 7 and I didn't get any errors. So I don't think its any CPU related error.

I got new set of logs using gdb.

http://pastebin.com/9WSy0f1A

Finally the issue is resolved. It was RAM. Yes, it was RAM. I didn't remove RAM on first hand because it didn't fail in memtest. Then after I removed RAM, I didn't see any segmentation fault error.

Thanks for the support guys.