4

I get an error that the system has a problem mounting / and do I want to Skip, Ignore, etc. I choose Ignore and usually everything is OK. I'm new to LINUX and Ubuntu. This machine will eventually be run headless to boot automatically when power is turned on.

How do I diagnose this or work around it?

4
  • check your /etc/fstab and make sure the UUID of each mount is the same UUID of each partition. use sudo blkid /dev/sda1 etc to check each partition. For me, sometimes the swap partition gets formatted when fiddling around with other distros, so has a different UUID.
    – duffydack
    Dec 31, 2011 at 22:59
  • @duffydack Please consider making this an answer to the question. Jan 6, 2012 at 6:05
  • Nothing caught my attention using blkid, alas. Jan 6, 2012 at 16:38
  • @duffydack so after issuing sudo blkid /dev/sda1...what exactly are we supposed to do?
    – puk
    Nov 15, 2013 at 4:24

3 Answers 3

4

I don't know how to diagnose graphically. Hopefully someone else will add an answer addressing that.

Is the problem linux related?

Take careful note of where you see the error. If it happens before the bootloader (grub), then what you see is probably a BIOS error, not linux related. You may find the rest of the steps useful, but probably not. If it happens at grub or after, then linux is reporting the error. Continue with this guide.

Search the boot messages

I would grep through dmesg to see any disk related errors. There are far better tutorials to get you up and running with linux commands, I recommend them over the following:

  • dmesg prints the kernel ring buffer, which near as I've been able to figure out is all the important messages about what linux is doing. If there was a boot problem, it ought to show up here.

  • grep is a search utility.

  • Hard disks in linux these days are named "sd"+letter. Ie, my first sata disk is named sda, my second sata disk is named sdb..

  • The pipe character, | creates a flow of information by pouring the output of the preceding program into the following program. (Some programs like grep use the pipe symbol internally to mean OR).

Put this together and you get:

dmesg | grep "sd[a-z]"
dmesg | grep "mount"
dmesg | grep -E "error|fail|warn"

Is there a problem post-boot?

Use fdisk and df to get printouts of your current disk situation.

  • sudo fdisk -l prints a list of disks and partitions.

  • df -h prints a list of mounted partitions and how much space is used vs available.

Does anything look weird? Are disks or partitions missing? Mis-sized? Do these programs report any errors?

7
  • @Marty Nothing stands out from all your suggestions. I get that there's an error on /. (not sure if the period is the end of the sentence or not) and the system waits for me to press f, or i, or ??. I end up with the boot choice screen and then everything appears to be OK. How should check the disk and how long might it take? It's only 8 GB? (If that might help.) Jan 5, 2012 at 22:22
  • @user75062 (1) Can you snap a pic of the problem and edit it into your question? (2) What do you mean by "end up with the boot choice screen", are you talking about grub?
    – djeikyb
    Jan 5, 2012 at 22:32
  • Yes, I should have said grub. Can't get a screen shot before boot. Jan 6, 2012 at 16:40
  • @user75062 So I understand, the order of events is (1) turn on computer (2) bios posts (3) error occurs, requires intervention (4) grub loads (5) ubuntu loads with everything apparently working. Is this correct?
    – djeikyb
    Jan 7, 2012 at 10:51
  • Yes, but the problem has gone away on its own - see my post above. Jan 7, 2012 at 18:55
1

The problem is that the computer had no battery for the hardware clock so there was a mismatch on times even though it did get the time via the network. I had loaded another Linux distribution on one of my other machines and it gave me an error message enough information to allow me to figure out how to check this. I've not had the problem since I installed batteries in 3 of the machines.

0

Whenever I've gotten that message, I found some error in /etc/fstab.

For example, I started getting an error when I upgraded to 11.10 that I didn't get previously, when mounting one of my NTFS partitions; the last field (fs_passno) was set to "2", but since it was NTFS, and couldn't be checked, it needed to be set to "0". I went to the trouble of booting the Windows XP install on the drive to check the partitions for errors, got the same error, then discovered the problem in fstab.

So, check fstab closely for the root partition entry. Run the command: sudo blkid -c /dev/null to get all the UUIDs, then check them carefully with fstab's entries to make sure there are no mistakes.

2
  • Here's what I get: Jan 5, 2012 at 22:19
  • Could find nothing wrong looking at fstab or using blkid. Jan 6, 2012 at 16:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .