LinuxQuestions.org - Debian Stable, crash experience: is there something else than EXT3?

- Debian (https://www.linuxquestions.org/questions/debian-26/)

- - Debian Stable, crash experience: is there something else than EXT3? (https://www.linuxquestions.org/questions/debian-26/debian-stable-crash-experience-is-there-something-else-than-ext3-639938/)

Debian Stable, crash experience: is there something else than EXT3?

Hi,

I am using Debian Stable since 1.5 years, and I am surprised It already occured 3 partition EXT3 big crashes (loss of big amount of data). Since I have been fervent user of windows, & this never happened with NTFS after 5 years due to regular chkdsk. Is there any other partition type better than this EXT3 for Linux ?
I think it s due to tune2fs that does not check else than / - mounted partitions.

Quote:

Originally Posted by frenchn00b (Post 3142962)

btw I just had a major NTFS problem over the weekend. Had to reformat due to lack of default repair tools!
Why are you using tune2fs? Is the journaling of ext3 not good enough? What happens when you just resort back to mounting your ext3 partitions as ext2 (thus without journaling enabled?). This is possible, right? Or am I missing something else here?

I can hardly imagine that another file system is going to make much of a difference. But what do you mean exactly by "crash"? Do partitions fail to mount? Do they suddenly become inaccessible while you are using them? I had issues of my own a while ago when one of my partitions would suddenly become unusable while it was being copied to. It turned out to be a simple matter of replacing a SATA cable. Problem solved.

Going on 5 years with my current system and never had an issue with EXT3.. I did however have a drive that started to fail, but since I was running smartmontools it notified me the drive was deteriorating early enough I was able to transfer my system to a new HD and not loose any data..

You might be having other hardware issues that are causing your problem, I highly doubt it's the file system. I agree with Jay73, you need to look for other issues that are more likely causing the problem.

Quote:

Originally Posted by farslayer (Post 3143379)

I tried using the tune2fs -c 2

Quote:
Originally Posted by makuyl View Post
Post "tune2fs -l /dev/sda1", /etc/e2fsck.conf and /etc/fstab.
Code:

Filesystem created: Wed Dec 13 22:19:08 2006
Last mount time: Sat May 10 12:05:02 2008
Last write time: Sat May 10 12:05:02 2008
Mount count: 19
Maximum mount count: 2
Last checked: Sun May 4 09:23:02 2008
Check interval: 15552000 (6 months)

tune2fs only fix the / first partition, and do not give interest to fixing others

--
I installed just now :

Code:

apt-get install smartmontools smart-notifier

I try to understand how it works ... and how to set up to my box

Quote:

Originally Posted by farslayer (Post 3143379)

If you have a quick command so that smartmontools fix, check and verify permanetntly a pc that runs all the time, no reboot, with 3 disks, please do not hesitate !

Quote:

Originally Posted by farslayer (Post 3143379)

I found an howto ... it 's complicated.. can it fix hte mounted disks ?
http://www.howtoforge.com/checking-h...-debian-ubuntu

welcome for any hints

smartmon tools is for testing and monitoring your hard drives for failure or pre-fail conditions. if your drive is generating errors of any kind the system can notify you before they become serious enough that you loose data.

Could always check the man page for more info...

Code:

EXAMPLES

      smartctl -a /dev/hda

      Print all SMART information for drive /dev/hda (Primary Master).



      smartctl -s off /dev/hdd

      Disable SMART on drive /dev/hdd (Secondary Slave).



      smartctl --smart=on --offlineauto=on --saveauto=on /dev/hda

      Enable  SMART on drive /dev/hda, enable automatic offline testing every

      four hours, and enable autosaving of SMART Attributes.  This is a  good

      start-up line for your system´s init files.  You can issue this command

      on a running system.



      smartctl -t long /dev/hdc

      Begin an extended self-test of drive /dev/hdc.  You can issue this com‐

      mand on a running system.  The results can be seen in the self-test log

      visible with the ´-l selftest´ option after it has completed.



      smartctl -s on -t offline /dev/hda

      Enable SMART on the disk, and begin an immediate offline test of  drive

      /dev/hda.  You can issue this command on a running system.  The results

      are only used to update the SMART Attributes,  visible  with  the  ´-A´

      option.  If any device errors occur, they are logged to the SMART error

      log, which can be seen with the ´-l error´ option.



      smartctl -A -v 9,minutes /dev/hda

      Shows the vendor Attributes, when the disk  stores  its  power-on  time

      internally in minutes rather than hours.



      smartctl -q errorsonly -H -l selftest /dev/hda

      Produces  output only if the device returns failing SMART status, or if

      some of the logged self-tests ended with errors.



      smartctl -q silent -a /dev/hda

      Examine all SMART data for device /dev/hda, but produce no printed out‐

      put.  You must use the exit status (the $?  shell variable) to learn if

      any Attributes are out of bound, if the SMART  status  is  failing,  if

      there  are errors recorded in the self-test log, or if there are errors

      recorded in the disk error log.



      smartctl -a -d 3ware,0 /dev/sda

      Examine all SMART data for the first ATA disk connected to a 3ware RAID

      controller card.

Code:

REFERENCES FOR SMART

      An introductory article about smartmontools is  Monitoring  Hard  Disks

      with  SMART,  by Bruce Allen, Linux Journal, January 2004, pages 74-77.

      This is http://www.linuxjournal.com/article.php?sid=6983 online.



      If you would like to understand better how SMART  works,  and  what  it

      does,  a good place to start is with Sections 4.8 and 6.54 of the first

      volume of the ´AT Attachment  with  Packet  Interface-7´  (ATA/ATAPI-7)

      specification.  This documents the SMART functionality which the smart‐

      montools utilities provide access to.  You can find Revision 4b of this

      document  at  http://www.t13.org/docs2004/d1532v1r4b-ATA-ATAPI-7.pdf  .

      Earlier and later versions of this Specification are available from the

      T13 web site http://www.t13.org/ .



      The  functioning of SMART was originally defined by the SFF-8035i revi‐

      sion 2 and the SFF-8055i revision 1.4 specifications.  These are publi‐

      cations of the Small Form Factors (SFF) Committee.  Links to these doc‐

      uments may be found in the References section of the smartmontools home

      page at http://smartmontools.sourceforge.net/ .

as for your current problem, I do not know the resolution to your issue.

Quote:

Originally Posted by farslayer (Post 3150466)

Thank you very much I will study this during the 2 months.I will take time to make it, since I am not specialist in informatic. I will temporally solve the problem with once a week setup at boot in the /etc/rc2.d/ a script to fsck.ext3 my harddisks.

I would do temporally:

/etc/rc2.d/SXXmyfscksaver :

Code:

#!/bin/sh

for each in "$(cat /etc/fstab | grep ext3 | awk ' { print $1 }  ')" ; do 

fsck.ext3 $each

done

Would you know which values of XX should I use for the disk fixer to be done at every boots, before the linux mounts all ?

tune2fs does not do other harddisks else than /. That's a very bad choice, since newbies can get peaced of debian (like I did, since 2 years ago I didnt know that the disks were very never checked by the debian stable box) I got 238 times mounted without any single fsck.ext3, imagine !

Quote:

Originally Posted by frenchn00b (Post 3150610)

tune2fs does not do other harddisks else than /. That's a very bad choice, since newbies can get peaced of debian (like I did, since 2 years ago I didnt know that the disks were very never checked by the debian stable box) I got 238 times mounted without any single fsck.ext3, imagine !

tune2fs doesn't check or fix anything. It is for _tuning_ purposes and as it doesn't know anything about the purpose of the underlying filesystem or it's mountpoint it's clearly able to change any ext2/3 partition.

in order to have your partitions checked at boot time, place a file called "forcefsck" in / and reboot. which partitions are checked then is determined by the second last column in /etc/fstab.

please read the manual pages of e2fsck, fstab and tune2fs.