Hello there,

i have a debian wheezy small box (Linux box 3.2.0-4-amd64 #1 SMP Debian 3.2.73-2+deb7u3 x86_64)

I'm newbie with linux, all i have done on this box was after a lot of hours reading, and fucking with vm's for months

So, its a small box, that i just run an ftp server and log bw usage with vnstat.

Everything been running smooth for an year and half until last weekend. I have a serious power outage, and psu died. Put a brand new on it, restarted and everything went fine, updated/upgraded, ftp was up and running, but ive noticed a problem with vnstat.

user@box:~$ vnstat
Database updated: Mon Mar 7 21:45:20 2016

Segmentation fault
user@box:~$


Ive did a lot of research on google, been on #debian official mirc channel ofr the last two days, but still can't figure it out.

After a lot of questions there, lemme explain what i did.

I've been running memtestx86 v4.20 for 10 hours on each memory slot, alone, not both slots ocuupied, and get no errors.

So, they told me to try an command called gdb. after a lot of time trying to run it properly, i managed to get the uotput they were requesting me about the vnstat alone.

here is the result with a backtrace included.

http://pastebin.com/XK6Q6Rgu

So, according to that log, seems it's an issue with "libc.so.6" or something like, i really don't understead.

I really need keep those logs from vnstat, if fix its a long way, its ok. Maybe we could just "uninstall" it and reinstall, than export logs back ?

If this is the best option, can anyone guide me how ? i have an second HD on that box (sdb) so i could backup vnstat data to there, do the process and later recover the logs.

Any help will be really apreciated.

Also, if i have to post any more info about the box or output from any command to help get more info to fix it, plz let me know.

thank you

Your vnstat databases ought to be in /var/lib/vnstat/ (they're just files, named after your interfaces). Try copying them to another machine and see if they can be read with vnstat there - if so, this is a problem with this machine, otherwise it's a problem with the databases themselves (which might have been corrupted when the machine lost power) and they'll need to be restored from backups.

CTM, first of all thank you for the answer.

Ok, i'll setup an debian vm on my windows desktop and try what you said, i'll be back with results.

thank you again

Do you have a clean wheezy sources.list?
Replacing libc.so.6 with a foreign or "updated" one can cause many problems, including segfaulting apps that call it expecting the "proper" version.

There are many clueless 'guides' about that suggest doing this, to get some package installed on (old/outdated/stable) Debian.

Hi again CTM and descendant_command,

sorry for the long time to answer, i had a very busy week.

well, in the end i tryied the copy the vnstat db to a vm to test it, and it was really corrupted, i couldn't recover any data, cause i didn't had an backup.

Anyway, ive properly uninstalled it and reinstalled, and its working fine again.


thank you both for the help