System wants to check drives on every boot

[MDuff]

I have a multi-boot setup (5 OS types) that has worked fine since installing a new HD 4 months ago but has just begun checking the drives at every boot. This occurs no matter which buntu or Mint OS I boot to.

I let the drive checks complete but the next time I boot the system goes through the entire procedure.

The only thing that has changed in the set up recently was the installation of Kubuntu 12.10 about a week ago but initially it worked OK as well.
It may have reached the 32 boot when I would expect a drive check but perhaps is not re-setting the counter or something.

I would appreciate any light that can be shed on this. Thanks

[james147]

I would try booting a livecd and running fsck manually.

[tek_heretik]

That's weird. Are all the OS file systems the same? Which OS is calling the check? If all 5 OSes are on the same drive, the first primary partition should only have the boot flag. Is the swap area shared? You have a pretty crowded HD, this could be caused by just about anything. I am guessing the drive is SATA, check the cable, seriously, a loose cable could cause write errors, is it mastered on any of the channels? Wow, I could go on for days, lol.

Somebody correct me if I'm wrong but is it possible 'crosswriting' (booting in to an OS and writing to partitions owned by another OS) can cause checks?

[james147]

Somebody correct me if I'm wrong but is it possible 'crosswriting' (booting in to an OS and writing to partitions owned by another OS) can cause checks?

Unlikely.

ext stores some information in it as to when a next check should be preformed, this is normally every 30 mounts or 180 days (which ever comes first) to me it sounds like this has triggered but is not being reset correctly, hence why I suggested to manually run the fsck so you can see what's going on.

[MDuff]

I would try booting a livecd and running fsck manually.

Excellent Idea! Should have thought of that myself.
In response to the other comments, the hardware is a laptop so not likely a cableing issue. The drive is configured with a Primary and Secondary partitions. The Primary has one OS installed and the secondary has four. All of them are about 12 Mb.

I will try fsck from a live CD. Thanks for the suggestion

[tek_heretik]

All of them are about 12 Mb.

Did you mean GB? 12Mb(its) is about 1.5MB(ytes)

[MDuff]

Did you mean GB? 12Mb(its) is about 1.5MB(ytes)

My BAD. Too much rush...... too little thought.

Just tried fsck from a live CD but no joy. It did tell me that the drive had not been checked in 46 starts and ran the entire check but when I tried to log in normally, the drive test ran again.

This laptop has so many variations of Linux because I mostly use it to show prospective Linux users some of the many faces of Linux to help them make a decison about which one they might like to install. It has had so many nix systems installed that it is probably about time I re-formatted the entire drive and rebuilt it.

I doubt that this problem will occur to anyone else. I can't mark it solved because it isn't. Someone might someday find an answer and post it in case it every crops up again.

I will post a followup when I rebuild the drive to put 30 to the story.

[dibl]

It might be worth a try to use the -pv options on a fsck from a Live CD or USB stick, to see if there are any messages about the filesystem. In a terminal:

 sudo e2fsck -pv /dev/sda1

(for whichever partition you have the troublesome OS on)

See if there are any messages about time stamps, bad blocks, etc.

[oshunluvr]

Check to see if a /forcefsck file exists on your system

[MDuff]

I cleared the drive, set up 3 primary partitions, 3 logical partitions in one of the primaries and installed two 'buntus. System runs fine as a dual boot but I will not install any more Linux until it has successfully completed the first drive check procedure.

I'll file your suggestion away dibl, just in case the drive check goes bad again. Thanks

[SteveRiley]

ext stores some information in it as to when a next check should be preformed, this is normally every 30 mounts or 180 days (which ever comes first)

Hm, hasn't this been disabled by default for some time now?

steve@t520:~$ [B]sudo tune2fs -l /dev/sda2 | egrep -i 'maximum mount count|check interval'[/B]
Maximum mount count:      -1
Check interval:           0 (<none>)

[james147]

Hm, hasn't this been disabled by default for some time now?

steve@t520:~$ [B]sudo tune2fs -l /dev/sda2 | egrep -i 'maximum mount count|check interval'[/B]
Maximum mount count:      -1
Check interval:           0 (<none>)

Didnt know that :P wonder why they did that...

[SteveRiley]

Didnt know that :P wonder why they did that...

Perhaps because modern file systems are sufficiently robust that scheduled checks are unnecessary now. I've seen the kernel pause boot and run fsck when it detects that a repair is needed, which is a better approach, IMHO.

[MDuff]

I finally found the problem.... or not. I usually create a directory for root, one for /home and of course one for swap.
If the root directory is formatted ext2 the computer will always, in my experience with two of computers, do a system check on startup.
If the root directory is formatted ext3 or 4 the computer will only do a system check about every 32 startups.
Strange that I could not find any references to this on my searches.

[SteveRiley]

EXT2 has no journal, so a file system check on startup makes sense. EXT3 and EXT4 have journals, which for the most part eliminate the requirement. Modern distros have skipped startup checks of these filesystem types completely.