LinuxQuestions.org
Share your knowledge at the LQ Wiki.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Software
User Name
Password
Linux - Software This forum is for Software issues.
Having a problem installing a new program? Want to know which application is best for the job? Post your question in this forum.

Notices


Reply
  Search this Thread
Old 05-28-2004, 01:32 AM   #1
haimeltjnfg
Member
 
Registered: May 2004
Distribution: debian
Posts: 235

Rep: Reputation: 30
Looking for file system that won't corrupt files on hard reboot


Ok - here's the story. I've been using debian sarge for the past month or so (probably a bit longer), and every time my computer locks up (to the point where i have to use the reset button), i'm unable to use part of my system. One type apt-get wouldn't work, another time kde wouldn't load, another time, X would freeze before the mouse cursor became visible. From what i gather, this is caused by file corruption when i press that evil button of doom labeled 'reset.' The only 2 file system's i've used were Ext3 and reiserfs. I've tried compiling with Sysrq, but it doesn't seem to work for me (yes, cat /proc/sys/kernel/sysrq shows 1).

anyway, i could use a nice file system that won't corrupt so easily (or that can take constant hits from the reset button). or i could use some utilities to run to make sure everything is OK. Hey, you know what, my hard drive might be going bad, is there a program i could use to check it?

is there a better distrubition i should try? (i'm trying gentoo, but the installition period of a couple days isn't too appealing)

oh here's my specs:

amd athlon xp 2200+
ecs k7s5a
nvidia geforce 3 (with nvidia-kernel, glx)
256 mb cas 2.5 @ 2 crucial ram pc2100
256 mb cas 3 @ 2 kingston ram pc2100
maxtor diamondmax 9 60 gb 7200 rpm hard drive:
/dev/hda1 50mb /boot
/dev/hda2 1gb swap
/dev/hda3 20gb /mnt/windows (fat32 parition)
/dev/hda4 extended
/dev/hda5 8 gb /home
/dev/hda6 1 gb /tmp
/dev/hda7 15 gb /usr
/dev/hda8 15 gb /

would mounting /usr with the ro attribute help? the problem is, i haven't ever been able to get to the point where i want to lock the /usr partition (i'm always finding more toys to play with)


case temp 101 F (i don't like noise in my case!)
cpu temp 120 F
hard drive 45 C ( (1.8 x 45) + 32 = 113 F ) or maybe it was 55 C. (131 F) uh oh. i think my hd is too hot. I have it suspended by bungee cords to quiet it down, but that makes it run hotter because heat can't dissapte through the metal of the case (but it can't resonate through the metal either)
 
Old 05-28-2004, 01:57 AM   #2
pjbgravely
Member
 
Registered: Mar 2004
Location: USA
Distribution: Aptosid, Ubuntu server, DSL
Posts: 53

Rep: Reputation: 16
Any file system will take damage in a hard shutdown, but why are you reseting? The only time I had that problem was when a bad driver that caused the keyboard to lock up. I was usually able to shutdown with the mouse but one of the times I couldn't I had to try to shut off power between hard drive activity. I missed, but that is what backups are for. If you have a working keyboard hit alt/ctrl/back space, this will drop out of X so you can log back in with KDM or GDM or take you to a shell where you will login and startX or Halt to shut down.
 
Old 05-28-2004, 02:02 AM   #3
haimeltjnfg
Member
 
Registered: May 2004
Distribution: debian
Posts: 235

Original Poster
Rep: Reputation: 30
Quote:
Originally posted by pjbgravely
Any file system will take damage in a hard shutdown, but why are you reseting? The only time I had that problem was when a bad driver that caused the keyboard to lock up. I was usually able to shutdown with the mouse but one of the times I couldn't I had to try to shut off power between hard drive activity. I missed, but that is what backups are for. If you have a working keyboard hit alt/ctrl/back space, this will drop out of X so you can log back in with KDM or GDM or take you to a shell where you will login and startX or Halt to shut down.
well the whole system freezes. i try hitting alt ctrl backspace 5 times over a span of a minute, and then i try hitting alt ctrl delete 5 times over a span of a minute. then i try alt sysrq S alt sysrq U and alt sysrq B, with no luck. i then am forced to give in and press reset. :-/ what filesystem will take the least damage?

do you think there is a possible hardware problem? this same computer with a hotter running athlon xp 1600+ ran rock solid in windoze.
 
Old 05-28-2004, 02:11 AM   #4
quatsch
LQ Addict
 
Registered: Aug 2003
Location: New York, NY
Distribution: gentoo, gentooPPC
Posts: 1,661

Rep: Reputation: 48
are there particular circumstances when your system locks up solid? On my laptop I get it from time to time when playing video using the Xv extension (xine, mplayer).

Also, after a hard reboot, are you told that the system did not shutdown properly and asked if you wanna run a file system check? I found that when I let it do a file system check AND let it 'repair' it, I',m likely to end up with a non-working system. Nowadays, I just let it boot without fschecking and it's usually fine. I use ext3 and I do make full system backups from time to time just in case.

Last edited by quatsch; 05-28-2004 at 02:22 AM.
 
Old 05-28-2004, 02:17 AM   #5
darthtux
Senior Member
 
Registered: Dec 2001
Location: 35.7480 N, 95.3690 W
Distribution: Debian, Gentoo, Red Hat, Solaris
Posts: 2,070

Rep: Reputation: 47
You probably want a journaling file system such as ext3 or ReiserFS.

Your running Sarge. Have you done an upgrade lately?
 
Old 05-28-2004, 07:15 PM   #6
haimeltjnfg
Member
 
Registered: May 2004
Distribution: debian
Posts: 235

Original Poster
Rep: Reputation: 30
i've tried ext3 and reiserfs. i run apt-get update && apt-get upgrade daily. :-/ It happens in X (kde 3.2.2) with juk running. The sound gets stuck in a short loop. How do i set it to boot without fschecking?
 
Old 05-28-2004, 07:40 PM   #7
MS3FGX
LQ Guru
 
Registered: Jan 2004
Location: NJ, USA
Distribution: Slackware, Debian
Posts: 5,852

Rep: Reputation: 360Reputation: 360Reputation: 360Reputation: 360
Running a machine on a file system that you know to be corrupted and not running any sort of repairs on it would be very stupid, to put it mildly.
 
Old 05-28-2004, 08:46 PM   #8
quatsch
LQ Addict
 
Registered: Aug 2003
Location: New York, NY
Distribution: gentoo, gentooPPC
Posts: 1,661

Rep: Reputation: 48
I know it sounds dumb not to run repairs but, trust me, I get better results when NOT repairing (it's ext3).

anyhow, you should be asked if the system thinks a file system check is needed even if it is not scheduled (ext3 is set to run an fsck from time to time).
 
Old 05-28-2004, 11:36 PM   #9
haimeltjnfg
Member
 
Registered: May 2004
Distribution: debian
Posts: 235

Original Poster
Rep: Reputation: 30
Quote:
Originally posted by quatsch
anyhow, you should be asked if the system thinks a file system check is needed even if it is not scheduled (ext3 is set to run an fsck from time to time). [/B]
it's never asked me.
 
Old 05-29-2004, 03:13 AM   #10
Electro
LQ Guru
 
Registered: Jan 2002
Posts: 6,042

Rep: Reputation: Disabled
I use XFS for most of my partitions. It caches a lot more than other filesystems and in my opinion it is fast on my system. For XFS to work to the best it can, you need atleast 512 megabytes of RAM. When I had a problem about three months ago with a corrupted partition table that mplayer destroyed, all I did is recreate the partition table and ran xfs_repair. After a few seconds, I got my files back. Some files were in lost+found but every file was perfectly intact even the file that mplayer is making. I had problems with ReiserFS trying to bring it back from the dead. I had no success with corrupted ReiserFS partitions.

I had problems with my Maxtor drive and its going to be the last Maxtor drive that I'm going to own (I'm sorry Maxtor fans). I enable the no write error with hdparm so the Maxtor drive will not detect bad sectors. IMO, its best to make the filesystem to do this. Usually its hardcoded in filesystems already and they should be smart enough to detect bad sectors.

Every once in a while run sfdisk to print out your partition table information and do backups on files that you can not recreate.

BTW, haimeltjnfg, never mix memory brands because it creates an unstable system.
 
Old 05-29-2004, 11:50 AM   #11
haimeltjnfg
Member
 
Registered: May 2004
Distribution: debian
Posts: 235

Original Poster
Rep: Reputation: 30
I just reinstalled debian on clean a clean reiserfs formatted drive last night, and this morning, i turned it on, and the mouse wasn't working so x wouldn't load. i modprobed the drivers (and the mouse worked with gpm) and typed startx. X froze. i couldn't toggle the num lock or caps lock or anything like that. i tried alt ctrl backspace. alt ctrl delete. alt sysrq s (then alt sysrq u, alt sysrq b), and nothing worked. I decided to wait, so i took a shower, and came back. it was still frozen. I had no choice but the press the reset button (unless their is something else i could do that i don't know about) it restarted fine. x won't work cuz the mouse driver modules aren't loading, so i had to loud them manually. this time i restarted kdm instead of startx. i get into kde open up konquerer and resize the window. x freezes. i can't move anything, hitting caps lock or num lock will not change the status of their respective lights, alt ctrl del/ backspace or anything else won't work. it's been locked up for a good 7 minutes now. i guess i have no choice but to hit the reset switch.

arghh!

i'll list everything i have:
ecs k7s5a
axp 2200+
256 crucical cas 2.5 @ 2
256 kingston cas 3 @2
pioneer dvd-rom atapi drive (using scsi-emulation) hdd
non-name dvd-rw/cd-rw drive (using scsi-emulation) hdc
maxtor 60 gb 7200 rpm drive hda
logitech internet navigator keyboard hooked up by ps/2
microsoft wheel mouse optical by usb (needs hid, mousedev modules)
nvidia geforce 3 (using nvidia supplied modules for x - are these known to be unstable?)
sb live value using alsa's snd-emu10k1
winmodem (not used.)

this system has always been extremely stable (never, ever, ever crashed before)

Last edited by haimeltjnfg; 05-29-2004 at 12:14 PM.
 
Old 05-30-2004, 06:47 AM   #12
Electro
LQ Guru
 
Registered: Jan 2002
Posts: 6,042

Rep: Reputation: Disabled
Take out either the crucial or the kingston memory module. Next buy same brand and model memory. Then take out the Soundblaster LIVE! because it tends to take over your computer's bus. On many of my systems the Soundblaster LIVE! corrupts data. Since I put the Turtle Beach Santa Cruz in my computer, I'm not having any problems. Using a usb mouse is ok, but to make your mouse reliable use PS/2. I think your mouse comes with a USB to PS/2 converter.

The most computer problems comes from bad power supplies and noisy AC lines. This is what I read from IBM reports. Use either Enermax, Power and Cooling, Zalman, or Antec power supplies. Do not forget to buy an in-line UPS too.

BTW, you may want to dust the system with compressed air.
 
Old 05-30-2004, 11:11 AM   #13
haimeltjnfg
Member
 
Registered: May 2004
Distribution: debian
Posts: 235

Original Poster
Rep: Reputation: 30
Quote:
Originally posted by Electro
Take out either the crucial or the kingston memory module. Next buy same brand and model memory. Then take out the Soundblaster LIVE! because it tends to take over your computer's bus. On many of my systems the Soundblaster LIVE! corrupts data. Since I put the Turtle Beach Santa Cruz in my computer, I'm not having any problems. Using a usb mouse is ok, but to make your mouse reliable use PS/2. I think your mouse comes with a USB to PS/2 converter.

The most computer problems comes from bad power supplies and noisy AC lines. This is what I read from IBM reports. Use either Enermax, Power and Cooling, Zalman, or Antec power supplies. Do not forget to buy an in-line UPS too.

BTW, you may want to dust the system with compressed air.
i fried my motherboard mouse ps2 port, so i'm stuck with usb. i have an antec 300w power supply

VCore 1: +1.50 V (min = +1.42 V, max = +1.57 V)
VCore 2: +2.48 V (min = +2.40 V, max = +2.61 V)
+3.3V: +3.33 V (min = +3.14 V, max = +3.46 V)
+5V: +4.94 V (min = +4.74 V, max = +5.24 V)
+12V: +11.96 V (min = +11.40 V, max = +12.60 V)
-12V: -12.46 V (min = -12.60 V, max = -11.41 V) ALARM
-5V: -5.54 V (min = -5.21 V, max = -4.76 V) ALARM
Stdby: +5.07 V (min = +4.74 V, max = +5.24 V)
VBat: +2.03 V
fan1: 2556 RPM (min = 0 RPM, div = 8)
fan2: 0 RPM (min = 3013 RPM, div = 8) ALARM
Case Temp: +27 C (low = +15 C, high = +45 C) sensor = thermistor
CPU Temp: +36 C (low = +15 C, high = +45 C) sensor = thermistor


Anyway, i think the reason for the instablity is because of nvidia's drivers/mixed with an amd athlon processor. I'm using the open source drivers right now and i haven't had a crash since i switched to nv (and i've tried to provoke one with script-fu in gimp (9000 pixel logo anyone?))

can somebody explain how to run xfs_repair, etc? i know you can run mount -o remount,ro /usr to remount the /usr partion as ro, but how can you unmount them?

if i get another crash i will try out what you recommended. thanks for the advice!

Last edited by haimeltjnfg; 05-30-2004 at 11:12 AM.
 
Old 05-30-2004, 08:19 PM   #14
pjbgravely
Member
 
Registered: Mar 2004
Location: USA
Distribution: Aptosid, Ubuntu server, DSL
Posts: 53

Rep: Reputation: 16
Crash recovery

I just has a crash on SuSE 9.0. I have two partitions for SuSE, / and /home both ReiserFS. I was deleting a 700MB ISO from my home partition and during the process the computer rebooted, I couldn't imagine what had happened, I was thinking I was rooted, and when SuSE came back up I wasn't surprised to see many kernel panics. I figured the kernel was messed because at every reboot I would get a different panic message. I booted into Mandrake 9.2 and tried to mount SuSE /, it mounted, I checked things out and it seemed ok. So I mounted SuSE home to get really recent backups when the computer rebooted while mounting. So i tried it again and the Mandrake locked up, so the next time I just used Mandrake to edit SuSE's fstab to remove home. I then I rebooted into SuSE as root and all was well except the home partition. I then ran fsck.reiser and fixed the partition, mounted it, copied it to another drive and then rebooted. Nothing was lost, it is a mystery why the computer rebooted because of a file system error. Also why did SuSE have a kernel panic when all that was wrong was a bad partition.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
UFS file system corrupt Tsuroerusu *BSD 3 11-22-2005 11:20 PM
Reiser file system / Hard Disk/ Hard Drive Problems Oxyacetylene Linux - Software 4 10-10-2005 03:24 PM
File test to make sure files aren't damaged/corrupt bluedevlx Linux - Newbie 1 11-04-2004 01:11 PM
File Test to make sure files aren't corrupt/damaged bluedevlx Programming 2 11-04-2004 12:50 PM
corrupt file system michaelc Linux - Newbie 1 12-14-2001 12:32 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Software

All times are GMT -5. The time now is 04:34 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration