LinuxQuestions.org
Support LQ: Use code LQ3 and save $3 on Domain Registration
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Newbie
User Name
Password
Linux - Newbie This Linux forum is for members that are new to Linux.
Just starting out and have a question? If it is not in the man pages or the how-to's this is the place!

Notices

Reply
 
Search this Thread
Old 01-11-2010, 08:09 PM   #1
Peterken
Member
 
Registered: Apr 2007
Location: Belgium
Distribution: Debian
Posts: 60

Rep: Reputation: 20
"Root has been mounted x times without checking. Check forced."


Hi,

I have seen this message more then once and I cannot find information that is understandable to me and which is recent, so maybe the situation has changed with newer kernels/file systems.

My question: is this normal behavior with linux and beside from increasing the number of start-ups after which this check runs, is there another way to avoid this?
Isn't there a way to do a check after the os has been loaded?

Thanks in advance for your answers!
Peter
 
Old 01-11-2010, 08:40 PM   #2
Drakeo
Senior Member
 
Registered: Jan 2008
Location: Urbana IL
Distribution: Slackware, Slacko,
Posts: 2,455
Blog Entries: 3

Rep: Reputation: 182Reputation: 182
well the program is called tune2fs. and it write a flag on the drive and the program fsck runs at run level 1 before the hard drives are mounted . fsck reads the tune2fs. but it is ok I have 3 hard drives with many partitions on it. So I understand how you feel. Never disable it. Unless you are very good.
linux has the tune set for so many mounts then it scans the drive partition for errors.
so sit back let it do it's job.
or as root or sudo type man tune2fs and set the count.
Quote:
tune2fs -c100 /dev/sda1 for one hundred mounts on sda1
Quote:
tune2fs -i0 /dev/sda1 disable on sda1 drive

Read this.
 
1 members found this post helpful.
Old 01-12-2010, 12:50 AM   #3
linuxlover.chaitanya
Senior Member
 
Registered: Apr 2008
Location: Nagpur, India
Distribution: Cent OS 5/6, Ubuntu Server 10.04
Posts: 4,627

Rep: Reputation: Disabled
Though it may be annoying sometimes if you have a lot of drives, I find this very helpful and useful. It does your job automatically at scheduled intervals that you might not do or tend to forget.
 
Old 01-13-2010, 08:34 PM   #4
Peterken
Member
 
Registered: Apr 2007
Location: Belgium
Distribution: Debian
Posts: 60

Original Poster
Rep: Reputation: 20
Hi guys, thanks for your input but sorry to say that I didn't get much wiser...

I have read the article that Drakeo linked to, and I understand some of it. Maybe the answers to a few questions will clarify things
1) does EVERY linux system do this check every +- 32 mounts and what about servers that are running 24/7?
2) what is the impact of tunefs on this behavior (checking every x times)
3) is it not possible to do a filesystem check by scheduling it after a machine has booted?

I have turned an older pc into a fileserver, mainly for backup and storage of my multimedia files with debian. It has 2 big disks 1TB and 1,5TB and I have everything in 1 partition so when it does a check, it takes quite some time. So I guess by partitioning it, this check would be finished faster, right?

You see, my questions are overlapping each other a little but I thought by posting them anyway, things might become more clear to me.

Thanks again for reading/replying!

P.
 
Old 01-13-2010, 08:44 PM   #5
Drakeo
Senior Member
 
Registered: Jan 2008
Location: Urbana IL
Distribution: Slackware, Slacko,
Posts: 2,455
Blog Entries: 3

Rep: Reputation: 182Reputation: 182
I guess I will do your school work lol.
Quote:
1) does EVERY linux system do this check every +- 32 mounts and what about servers that are running 24/7?
No it is different per operating system .
Quote:
2) what is the impact of tunefs on this behavior (checking every x times)
A sector goes bad and they do but you have millions more this allows the system to move the data on a bad sector. and not use that sector.
Quote:
3) is it not possible to do a file system check by scheduling it after a machine has booted?
yes if you have multi partitions you must umount the/dev then run fsck /dev/you want.
but if you have a device that that has your system on it and you need to to run the computer from that device.
you must boot to single mode telinit 1 that is in read only mode and the run the e2fsck /dev/yours.

Last edited by Drakeo; 01-13-2010 at 08:49 PM.
 
1 members found this post helpful.
Old 01-13-2010, 08:53 PM   #6
Elv13
Member
 
Registered: Apr 2006
Location: Montreal,Quebec
Distribution: Gentoo
Posts: 825

Rep: Reputation: 128Reputation: 128
1) No, some disable it or add the possibility to skip it by pressing escape when it run (like Ubuntu). It is a good practice to let it on. Windows 95-XP do that too when you are on a FAT32 partition because it is not a stable system. Windows on NTFS have some basic self healing mechanism but the good and old scandisk is there when you need it. Ext* is more stable than most file system, but good practices are good practices, it is always better to heal a file system, not only tolerate and recover from error (it do both). Ext is made to never fully break, it have advanced tree recovering, super block redondency, journal buffering and self detectable "need a real fsck" trigger.

2) You take more risks, if you like to unplug the A/C from your computer when it is on, hard resetting when the drive is spinning and do stupid things like that, you will end up with lost data sooner than when you keep the FS in perfect shape. See that like this: when you broke half of something that is already half broken, you end up with a totally broken toy but if you break half of a perfectly working on, the other half still work.

3) Yes it is,
Code:
touch /forcefsck
That command add a file on the root directory of the file system, if this file is present at boot, then scan will be ran, even if the FS self monitoring feature say it is fine.
 
1 members found this post helpful.
Old 01-13-2010, 08:55 PM   #7
Quakeboy02
Senior Member
 
Registered: Nov 2006
Distribution: Debian Squeeze 2.6.32.9 SMP AMD64
Posts: 3,238

Rep: Reputation: 121Reputation: 121
Quote:
1) does EVERY linux system do this check every +- 32 mounts and what about servers that are running 24/7?
Quote:
No it is different per operating system .
And it is different for each filesystem type. For instance, the XFS doesn't have a filesystem check, and fsck.xfs simply returns a zero (success), if there is an fsck.xfs on your system.
 
1 members found this post helpful.
Old 01-14-2010, 08:02 AM   #8
Peterken
Member
 
Registered: Apr 2007
Location: Belgium
Distribution: Debian
Posts: 60

Original Poster
Rep: Reputation: 20
@Drakeo (thank you schoolmaster! )
That's why I was puzzled because I was in course some time ago where they were using ubuntu 9.x and I had not seen this behavior there.
I understand that sectors go bad and that you cannot check a disk completely when an os is running from it, but since not all distros do this checking and it's done only on linux, I was even more 'in the dark'.

@Elv13
From what I understand, creating forcefsck will only force a check a next reboot (which the filename already implies) but that doesn't affect the normal checking interval.

How come ext3 'needs' this checking? I mean, it seems as if ext3 is inferior because of this or am I seeing this wrong?
 
Old 01-14-2010, 08:18 AM   #9
jschiwal
Guru
 
Registered: Aug 2001
Location: Fargo, ND
Distribution: SuSE AMD64
Posts: 15,733

Rep: Reputation: 654Reputation: 654Reputation: 654Reputation: 654Reputation: 654Reputation: 654
Quote:
How come ext3 'needs' this checking? I mean, it seems as if ext3 is inferior because of this or am I seeing this wrong?
Because ext3 does this checking, it is a superior in terms of robustness. Other filesystems are faster, or optimized for large files, or highly random access such as database use. These filesystems tend to be used for servers that operate 24/7 so tracking the number of times it is mounted doesn't do anything.
 
1 members found this post helpful.
Old 01-14-2010, 08:24 AM   #10
Drakeo
Senior Member
 
Registered: Jan 2008
Location: Urbana IL
Distribution: Slackware, Slacko,
Posts: 2,455
Blog Entries: 3

Rep: Reputation: 182Reputation: 182
it is not the ext3 that is bad it is the rust on your harddrive platter.
Ubuntu does fsck -f /dev/yours that is when you get that serious long boot up screen it is done the fsck.
RUST yes rust we burn that format in and yes micro rust. This will be argued because the hard drive is built in clean rooms. But not in a vacuum. This is why we can enable smart drive in the bios. BIOS will read smart drive to see if there is a change in the image and this will also force fsck. when this happens it is time to get the data off that drive it is failing.

When a platter or or an arm that holds the head gets vibrated real hard can touch the platter an hurt the image sectors. fsck can replace the image and restore the data. If the data is corrupted or unable to place back on the image it will be coppied to the "lost" folder this is on ext3 and up. then after booting you can go to the /lost folder and look at the data and put it back on the drive. Were as ext2 or fat32 this is not as easy and you usually loose the data.

Last edited by Drakeo; 01-14-2010 at 08:30 AM.
 
1 members found this post helpful.
Old 01-14-2010, 09:09 AM   #11
Peterken
Member
 
Registered: Apr 2007
Location: Belgium
Distribution: Debian
Posts: 60

Original Poster
Rep: Reputation: 20
Quote:
Originally Posted by jschiwal View Post
... Other filesystems are faster, or optimized for ... or highly random access such as database use. These filesystems tend to be used for servers that operate 24/7 ...
What about a machine running 24/7 serving a database; is ext3 not the best choice there?

@Drakeo: you mean that ubuntu does a check? I can't recall having noticed anything.

Guys, if that check can only be done at bootup/unmounted drive: what about a server that is running 24/7?
Does it make any difference in terms of time needed to check/delay if a disk is partitioned and the root is only a part of a disk?

Thanks for your replies!
P.
 
Old 01-14-2010, 06:36 PM   #12
jschiwal
Guru
 
Registered: Aug 2001
Location: Fargo, ND
Distribution: SuSE AMD64
Posts: 15,733

Rep: Reputation: 654Reputation: 654Reputation: 654Reputation: 654Reputation: 654Reputation: 654
Quote:
Originally Posted by Peterken View Post
What about a machine running 24/7 serving a database; is ext3 not the best choice there?

@Drakeo: you mean that ubuntu does a check? I can't recall having noticed anything.

Guys, if that check can only be done at bootup/unmounted drive: what about a server that is running 24/7?
Does it make any difference in terms of time needed to check/delay if a disk is partitioned and the root is only a part of a disk?

Thanks for your replies!
P.
IIRC, JFS is tuned for database use. XFS is tuned for accessing large files. There are articles on the web comparing EXT3, ReiserFS, XFS and JFS. For the root partition, EXT3 would be fine.

Often the devil is in the details. For example, performance for a filesystem may vary if it is on a raid-5 array or LVM2 volulme. One really needs to test different filesystems for themselves before committing which one is best for them. For large commercial database, you may have an extemely large ram caching device with it's own UPS inside the server that is used by a caching service.
 
1 members found this post helpful.
Old 01-16-2010, 06:16 AM   #13
Peterken
Member
 
Registered: Apr 2007
Location: Belgium
Distribution: Debian
Posts: 60

Original Poster
Rep: Reputation: 20
So choosing a filesystem on a linux machine is not that straightforward and apparently you have to weigh the pros and cons for your purpose.

In the beginning of the boot messages, I noticed ext2 as a file system but the fstab states it's ext3. Does this mean that there is a 'switch/conversion' every time?

Does it make any difference in terms of time needed to check/delay if a disk is partitioned and the root is only a part of a disk?
 
Old 01-17-2010, 10:03 PM   #14
Drakeo
Senior Member
 
Registered: Jan 2008
Location: Urbana IL
Distribution: Slackware, Slacko,
Posts: 2,455
Blog Entries: 3

Rep: Reputation: 182Reputation: 182
Quote:
What about a machine running 24/7 serving a database; is ext3 not the best choice there?

@Drakeo: you mean that ubuntu does a check? I can't recall having noticed anything.
I have been troble shooting Ubuntu for a long time. and yes it does fsck at boot.
like my server my stuff is backed up all the time. when you are running many partitions and many drives. Then you will find it would be nice

umount /dev/sd-? fsck -f -y /dev/sd-?
You are the keeper of your universe it is up to you to set your system up with redundant raid drives. so your stuff is back up and the server can keep going. Hot swap drives been around along time.
Back up back up back up. and in the server world back up to another server. in case yours goes down that one is up.
if your not smart enough to build your system and your cgi to work when drives are down then you will never have a server run 24 hours.
 
1 members found this post helpful.
Old 01-18-2010, 06:43 PM   #15
Peterken
Member
 
Registered: Apr 2007
Location: Belgium
Distribution: Debian
Posts: 60

Original Poster
Rep: Reputation: 20
Maybe it's because ubuntu hides the boot messages and those systems (laptop's with already not that big disks) where partitioned that I didn't notice it...
Quote:
Originally Posted by Drakeo
umount /dev/sd-? fsck -f -y /dev/sd-?
So I guess that this is the way to handle a system that is running 24/7?

I know that backup/redundancy is an important issue, it happened to me twice - quite some time ago - so now my important data is being backed up from my pc towards my server every day & my music collection on my server is synchronized to a second disk in the server.

I'm facing a redesign of my server and there are multiple things to consider:
- centralized storage
- move my tv recording from a client to a server
- virtualisation
- extended firewalling for whole network
- backup/redundancy
so for now I will keep my server running with 2 disks, each with different data on it.
 
  


Reply

Tags
boot, check, ext2, ext3, fsck, linux, mounting, suse, tune2fs


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
When "not cleanly unmounted", why the "check forced" is not not applied at boot? frenchn00b Debian 1 09-06-2009 11:21 AM
Data transfer online is slow "at times" or "stops at times" Balarabay1 Suse/Novell 14 04-30-2006 10:00 AM
Mandrake 10.1 hangs at "Checking root file system" while booting ananya Mandriva 1 03-01-2006 06:03 PM
"forced check" when booting Joe Soap Linux - Newbie 6 10-10-2002 02:25 AM
"File system with errors, check forced" Luis Rosso Linux - General 5 02-02-2001 07:26 PM


All times are GMT -5. The time now is 08:56 AM.

Main Menu
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
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration