HowTo: Change disk checking/fsck at boot frequency

[Old Pink]

Hi,

Nearly all Ubuntu users know, that every 30 boots their system performs what is know as an "fsck" to make sure your hard drive has no errors.

If you go through periods of frequent rebooting, these checks can become annoying.

To change how often these boot up checks occur, run this command in a terminal:

sudo tune2fs -c 50 /dev/hda1The two things in bold can be modified.

First, the number "50", which chooses how often you want the check performed. 1 makes it scan at every boot, 0 stops scanning altogether, and all other numbers mean the scan will be performed every ** boots, 20 for 20, 30 for 30, 40 for 40 and so on.

Next, the "hda1" which you simply change to match the location of your hard drive, be it hda1,2,3 or even sda1,2,3.

Setting the scan to never happen isn't recommended, I'd say between 50-100 is safe. :)

- Matt

[bodhi.zazen]

Nice How-to Old Pink

This thread has been added to the UDSF wiki.

Change Disk Check Frequency (http://doc.gwos.org/index.php/ChangeDiskCheckFrequency)

I took the liberty of adding some basic information about fstab:

fstab options
/etc/fstab is a system configuration file and is used to tell the Linux kernel which partitions (file systems) to mount and where on the file system tree.

A typical fatab entry may look like this:
/dev/hda1 / ext3 defaults 1 1
/dev/hdb1 /home ext3 defaults 1 2

The 6th column (in bold) is a fsck options.
0 = Do not check.
1 = First file system (partition) to check;
/ (root partition) should be set to 1.
2 = ALL OTHER file systems to be checked.

bodhi.zazen

[Charles Hand]

Short of turning it off completely (not recommended) it's still going to happen by surprise and on the computer's timetable. Is there any way to tie into the mechanism that counts the boots and run a little program that says "you're going to have a forced check three boots from now"? Maybe a log file that tells how may boots are left, which could be grepped?

[Sam]

There is also the -i switch to set an interval before the check.

-i interval-between-checks[d|m|w]
Example combining both methods:

$ sudo tune2fs -c 50 -i 1m /dev/hda1

[Charles Hand]

Short of turning it off completely (not recommended) it's still going to happen by surprise and on the computer's timetable. Is there any way to tie into the mechanism that counts the boots and run a little program that says "you're going to have a forced check three boots from now"? Maybe a log file that tells how may boots are left, which could be grepped?

Here it is:

sudo dumpe2fs -h /dev/sdb1 | grep -i 'mount count'

Output:


dumpe2fs 1.38 (30-Jun-2005)
Mount count: 5
Maximum mount count: 30

[kvonb]

Search the forums for "bonager", it is a panel app that warns you of impending scans and lets you postpone them :)

[MaineDave]

There's also a showfsck package (I'm using Edgy in case that matters) with the command "showfsck" to show how many reboots remain before a fsck runs.

[lunamystry]

how do i get the Ubuntuuser icon for my website?

[NuK]

I'm learning bash and I wrote this script that shows the "mount count" for each partition. It works fine for me... I hope it helps!


#!/bin/bash

# this directory must exist to execute the script
if [ -d /dev/disk/by-id/ ]; then :; else echo 'ERROR: /dev/disk/by-id/ does not exist'; exit 1; fi

# root user is necessary
if [ $UID -gt 0 ]; then echo 'ERROR: You have to be root'; exit 1; fi

# list ALL partitions and exclude the SWAP ones
declare -a SWAPS=( $(cat /proc/swaps | grep /dev/ | awk '{print$1}') )
lsParts="ls -l /dev/disk/by-id/ | grep --regexp=-part[0-9] | awk '{sub(\"../../\",\"/dev/\");"
for i in $(seq 0 $((${#SWAPS[@]} - 1))); do
lsParts=${lsParts}' sub("'${SWAPS[$i]}'", "");'
done
lsParts=${lsParts}" print\$10}'"
declare -a PARTITIONS=(`eval $lsParts`)

# format TEMP file
for i in $(seq 0 $((${#PARTITIONS[@]} - 1))); do
PARTITION=${PARTITIONS[$i]}
if [ -n $PARTITION ]; then
# first line, device name
echo $PARTITION >> .temp.$$
# mount info
sudo dumpe2fs -h $PARTITION | grep 'Filesystem volume name' >> .temp.$$
sudo dumpe2fs -h $PARTITION | grep 'Mount count' >> .temp.$$
sudo dumpe2fs -h $PARTITION | grep 'Maximum mount count' >> .temp.$$
sudo dumpe2fs -h $PARTITION | grep 'Last mount time' >> .temp.$$
sudo dumpe2fs -h $PARTITION | grep 'Last checked' >> .temp.$$
echo '-------------------------------' >> .temp.$$
fi
done

# show and delete TEMP file
clear
cat .temp.$$
rm -f .temp.$$

# TUNE help
echo
echo 'HOWTO'
echo -e "- Change Filesystem volume name:\n\tsudo tune2fs -L NewLabel /dev/hda1\n\tsudo e2label /dev/hda1 NewLabel"
echo
echo -e "- Change Maximum mount count:\n\tsudo tune2fs -c 50 /dev/hda1"
echo
echo -e "- Modify reserved space:\n\tsudo tune2fs -m 1 /dev/hda1"
echo

exit 0


Remember that you can call the script like this:

sudo bash ./scriptname.sh

But if you want to use it more than once, it's better to change the execute permissions and call it without the "bash" command:

chmod 755 ./scriptname.sh
sudo ./scriptname.sh

[ntlam]

Hi,

Nearly all Ubuntu users know, that every 30 boots their system performs what is know as an "fsck" to make sure your hard drive has no errors.

If you go through periods of frequent rebooting, these checks can become annoying.

To change how often these boot up checks occur, run this command in a terminal:

sudo tune2fs -c 50 /dev/hda1The two things in bold can be modified.

First, the number "50", which chooses how often you want the check performed. 1 makes it scan at every boot, 0 stops scanning altogether, and all other numbers mean the scan will be performed every ** boots, 20 for 20, 30 for 30, 40 for 40 and so on.

Next, the "hda1" which you simply change to match the location of your hard drive, be it hda1,2,3 or even sda1,2,3.

Setting the scan to never happen isn't recommended, I'd say between 50-100 is safe. :)

- Matt

Nice guide

[mallow]

OK, thanks for that. It`s helpfull. But is there a way to skip this check on startup? I man for example that I`m in a hurry and when I boot up ubuntu fsck starts. How to skip this and postpone to another boot? Is it possible or will it be in gutsy?

[TearDownAllSigns]

To skip on startup and move to shutdown you can use AutoFsck

https://wiki.ubuntu.com/AutoFsck

I haven't used it myself so I can't vouch but from the thread it sounds like it is working for people.

http://ubuntuforums.org/showthread.php?t=566058&highlight=fsck+on+startup&page=2

[ssam]

To skip on startup and move to shutdown you can use AutoFsck

https://wiki.ubuntu.com/AutoFsck


i am using this. its only run once so far (only had it installed a few weeks).

seem to work exactly as it says.

[j1solutions]

Hi,

Nearly all Ubuntu users know, that every 30 boots their system performs what is know as an "fsck" to make sure your hard drive has no errors.

If you go through periods of frequent rebooting, these checks can become annoying.

To change how often these boot up checks occur, run this command in a terminal:

sudo tune2fs -c 50 /dev/hda1The two things in bold can be modified.

First, the number "50", which chooses how often you want the check performed. 1 makes it scan at every boot, 0 stops scanning altogether, and all other numbers mean the scan will be performed every ** boots, 20 for 20, 30 for 30, 40 for 40 and so on.

Next, the "hda1" which you simply change to match the location of your hard drive, be it hda1,2,3 or even sda1,2,3.

Setting the scan to never happen isn't recommended, I'd say between 50-100 is safe. :)

- Matt

Thanks Old Pink, I know this is an old post, but I just found it and it's well explained!

[akhil.t]

I tried this command and it said that the frequency has been changed. But the fsck is still taking place every time I boot. If I do not press skip, the check stops and 50% and the system prompts me to do a manual fsck.

Any help? :-|

[scorp123]

the check stops and 50% and the system prompts me to do a manual fsck. Then I suggest you do as the system tells you because it's obvious you have a corrupted disk and are about to lose your data soon.

[akhil.t]

Then I suggest you do as the system tells you because it's obvious you have a corrupted disk and are about to lose your data soon.

Thank you scorp123. :)
I did that once. Ended up with a blank screen after almost 2 hrs of check. Had to reinstall the OS.

I am unable to understand why the command is not having the desired effect.

[bodhi.zazen]

You need to run that command from a live CD with the partition Unmounted.


fsck -rfv /dev/sdxy

where /dev/sdxy = the partition needing to be checked. Sometimes the problem can not be solved automatically in which case you may need to attempt manual recovery.

If automatic recovery fails, however, you usually have a serious hardware problem and you may need data recovery tools ...

How to fsck: http://www.linux.com/feature/32026

[akhil.t]

You need to run that command from a live CD with the partition Unmounted.


fsck -rfv /dev/sdxy

where /dev/sdxy = the partition needing to be checked. Sometimes the problem can not be solved automatically in which case you may need to attempt manual recovery.

If automatic recovery fails, however, you usually have a serious hardware problem and you may need data recovery tools ...

How to fsck: http://www.linux.com/feature/32026 Thank you bodhi. I'll try it out. :)

[directcharitycontribution]

Hi,

Nearly all Ubuntu users know, that every 30 boots their system performs what is know as an "fsck" to make sure your hard drive has no errors.

If you go through periods of frequent rebooting, these checks can become annoying.

To change how often these boot up checks occur, run this command in a terminal:

sudo tune2fs -c 50 /dev/hda1The two things in bold can be modified.

First, the number "50", which chooses how often you want the check performed. 1 makes it scan at every boot, 0 stops scanning altogether, and all other numbers mean the scan will be performed every ** boots, 20 for 20, 30 for 30, 40 for 40 and so on.

Next, the "hda1" which you simply change to match the location of your hard drive, be it hda1,2,3 or even sda1,2,3.

Setting the scan to never happen isn't recommended, I'd say between 50-100 is safe. :)

- Matt

http://manpages.ubuntu.com/ the options are shown with terminal > man fsck OR man tune2fs

and thread http://ubuntuforums.org/showthread.php?t=937345

[dannyman]

I boiled this down into a "universal" command:


sudo tune2fs -c 0 `mount | awk '$3 == "/" {print $1}'`

I posted further explanation on my web site:
http://dannyman.toldme.com/2008/11/19/e2fsck-frog-off/

[yangeryanger]

hi. i run a home server running 24/7 day and night and data is always moving from one hdd to another or some sorts.. is there a way to make it fsck more often? should i just lower the number in the sudo tune2fs -c 50 /dev/hda1 to sudo tune2fs -c 2 /dev/hda1 or something?

oh, edit: i'm using ext3 on the main boot, and on my other drives, jfs...

thanks guys.

[bodhi.zazen]

Well, you should not need to run it any more often, especially if you are moving files like that.

Also fsck should not be run on a mounted partition, so it runs at boot.

You should be able to run fsck at next boot if you like :

http://www.cyberciti.biz/faq/linux-force-fsck-on-the-next-reboot-or-boot-sequence/

[wamukota]

I boiled this down into a "universal" command:

sudo tune2fs -c 0 `mount | awk '$3 == "/" {print $1}'`


I don't use the 'c' (for count) between checks, but the 'i' (for interval).

So to set the interval every 2 weeks the command becomes:

sudo tune2fs -i 2w `mount | awk '$3 == "/" {print $1}'`

Every 10 days becomes

sudo tune2fs -i 10d `mount | awk '$3 == "/" {print $1}'`

Every 3 months becomes

sudo tune2fs -i 3m `mount | awk '$3 == "/" {print $1}'`

Using the interval allows you to predict more or less when the check will take place (ie every sunday, if you run the command on a sunday with an interval of 7 days)

[ciderpunx]

Is there any way to tie into the mechanism that counts the boots and run a little program that says "you're going to have a forced check three boots from now"? Maybe a log file that tells how may boots are left, which could be grepped?

Using the solutions already suggested you can do something like this:


$ dumpe2fs -h /dev/sda1 2>&1 | grep -i 'mount count' | perl -ne '$i++;chomp;s/.*://;s/\s//g;$arr[$i]=$_;print "Next fsck in " . ($arr[$i] - $arr[1]) . " mounts\n" if($i==2);'You can send that to a file by appending:

> /path/to/fileIf you have zenity installed, you can have a popup appear by appending this instead:

| zenity --text-info --title "fsck info"I have the last version saved as a shell script which can be set up to run in my startup applications (system|preferences|startup applications|startup programs). Here's that version, which is a bit tidier.


#!/bin/bash
# fsck mount dialogue hack, by ciderpunx#

dumpe2fs -h /dev/sda1 2>&1 \
| grep -i 'mount count' \
| perl -ne '$i++;
chomp;
s/.*://;
s/\s//g;
$arr[$i]=$_;
print "Next fsck in " . ($arr[$i] - $arr[1]) . " mounts\n" if($i==2);' \
| zenity --text-info --title "fsck info"

[ario]

Thanks. it saved my time a lot:p

[COKEDUDE]

Great thread :). I know its but I wanted to bring it back to life :).