0

While rebooting my Linux box today, I get a:

Read Error after the splash screen

And it doesn't boot. I think I have 14.04, on a recently upgraded 32-proc 64Gb machine. I've gone into the BIOS, and have reordered the boot drives; doing that just leads to a blinking cursor. I have 4 HDs, only the first is boot drive. I don't think any other HD is loaded to function as a boot drive.

Occasionally, it comes up with grub rescue> but I don't see what to do. If I type ls, I get the list of hd's I expect and I can see some files doing ls (hd0,msdos)/ though the list seems incomplete.

Philippe Delteil
  • 1,668
john Van Drie
  • 3
  • 1
    Lets test the safest way first... boot to a Ubuntu Live DVD/USB. Start the Disks application, select your boot disk, go to the "hamburger" icon and select SMART Data & Tests, screenshot the Data and post it to imgur.com for me. Then run the SMART tests. See my answer for how to run fsck. Report back and we'll do further tests. Start comments to me with @heynnema or I may miss them. – heynnema Apr 16 '19 at 15:01
  • That sounds like a disk error - can you boot from a live USB and execute fsck on the disks? Please refer to https://askubuntu.com/a/642789/283721 – Charles Green Apr 16 '19 at 15:02
  • @CharlesGreen please see my partial answer for a more concise way to fsck. – heynnema Apr 16 '19 at 15:06
  • @heynnema - Thanks! I had found the other answer just a couple of days ago. I seem to have seen several questions that related to disk errors recently. Do you mind if I expand upon your answer a little to include checking other partitions, such as a separate /home partition? – Charles Green Apr 16 '19 at 15:10
  • @CharlesGreen help yourself :-) You may wish to edit the 17.10 fsck paragraph also. – heynnema Apr 16 '19 at 15:39
  • Finally here's the fsck output sudo fsck -f /dev/sdf fsck from util-linux 2.31.1 e2fsck 1.44.1 (24-Mar-2018) ext2fs_open2: Bad magic number in super-block fsck.ext2: Superblock invalid, trying backup blocks... fsck.ext2: Bad magic number in super-block while trying to open /dev/sdf

    The superblock could not be read or does not describe a ...

     – john Van Drie Apr 16 '19 at 17:45
  • @johnVanDrie the fsck command's /dev should be a partition number... like /dev/sdf1 if that's your Ubuntu or /home partition(s). – heynnema Apr 16 '19 at 17:52
  • @heynnema. Aha. Here's using /dev/sdf1 sudo fsck -f /dev/sdf1 fsck from util-linux 2.31.1 e2fsck 1.44.1 (24-Mar-2018) ext2fs_open2: Bad magic number in super-block fsck.ext2: Superblock invalid, trying backup blocks... Superblock needs_recovery flag is clear, but journal has data. Recovery flag not set in backup superblock, so running journal anyway. /dev/sdf1: recovering journal Error reading block 34650768 (Input/output error). Ignore error? – john Van Drie Apr 16 '19 at 17:56
  • @heynnema I'm now trying e2fsck, after having run mke2fs -n /dev/sdf1, which lists a dozen superblock backups. I chose 229376, and am now running e2fsck -b 229376 /dev/sdf1. It's telling me it was not cleanly unmounted, check forced. Now it says Pass 1: Checking inodes, blocks, sizes. It said extent tree could be narrower, Fix [y]. I said yes. it's continuing – john Van Drie Apr 16 '19 at 18:04
  • @johnVanDrie lets try and keep the fsck comments under my fsck answer please. See my last comments there. – heynnema Apr 16 '19 at 18:11

1 Answers1

1

Lets first check your file system for errors.

For 17.10 or older...

  • boot to the GRUB menu
  • choose Advanced Options
  • choose Recovery mode
  • choose Root access
  • at the # prompt, type sudo fsck -f /
  • repeat the fsck command if there were errors
    • If you have your /home directories on a separate partition, run sudo fsck -f /home as well
  • type reboot

For 18.04 or newer...

  • boot to a Ubuntu Live DVD/USB
  • open a terminal window
  • type sudo fdisk -l
  • identify the /dev/XXXX device name for your "Linux Filesystem"
  • type sudo fsck -f /dev/XXXX # replacing XXXX with the number you found earlier
  • repeat the fsck command if there were errors
    • If you have separate partitions for your /home directory, repeat this using the appropriate /dev/xxxx
  • type reboot

Update #1:

The HDD is getting ECC correctable read errors, and failing fsck.

Update #2:

Data is being recovered from the drive. The drive will be replaced with a SSD.

heynnema
  • 70,711
  • I'm preparing a Ubuntu USB right now. In the past, I'd seen the GUI for the GRUB menu, but I'm not seeing that now. Instead I get a cmd line grub rescue> which doesn't appear to have disk-checking option. I'll let you know in 30 min or so how the Ubuntu USB works, I'm creating an 18.04-2 ISO – john Van Drie Apr 16 '19 at 15:36
  • @johnVanDrie Do the SMART stuff first, then the fsck. – heynnema Apr 16 '19 at 15:41
  • The SMART data shows all HD's: overall assessment is OK. Would the simplest thing be to install Ubuntu from the USB to one of the other HD's, and boot from there. Worry about the bad disk later/ – john Van Drie Apr 16 '19 at 16:29
  • @johnVanDrie No. Let finish troubleshooting first. Even though the SMART Data says OK, it may not be. That's why I want to look at it. Can you post the screenshot to imgur.com for me? Then do the fsck. Then we may look at syslog for more data. – heynnema Apr 16 '19 at 16:36
  • OK - thanks so much. While I try to get this s-shot, I did notice I'd overlooked the boot drive. When trying to mount in the Disk Utility tool, I get the ff. error: error mounting /dev/sdf1 at /media/ubuntu.... wrong fs type, bad option, bad superblock on /dev/sdf1, misssing codepage or helper program, or other err (udisks-error-quark, 0). It now says SELF-TEST FAILED. – john Van Drie Apr 16 '19 at 16:51
  • Screenshot just uploaded to imgur.com. It shows the output of the SMART run. – john Van Drie Apr 16 '19 at 17:00
  • I should add: even if I stop using this HD as a boot device, it has important data stored on it. I last did a backup a few weeks ago, and stuff since then I'd like to pull out of the disk, if possible. – john Van Drie Apr 16 '19 at 17:03
  • @johnVanDrie you have to tell me the imgur.com URL so I can go look at it. Edit your question and show me fdisk -l. Start comments to me with @heynnema or I may miss them. – heynnema Apr 16 '19 at 17:09
  • https://imgur.com/a/SJHlaqG? The output of fdisk -l is over 100 lines. Do you need just a portion of that? It looks like it's reporting stuff on all HD's. I've unmounted all but the USB and the bad one. – john Van Drie Apr 16 '19 at 17:17
  • https://imgur.com/a/SJHlaqG? The output of fdisk -l is over 100 lines. Do you need just a portion of that? It looks like it's reporting stuff on all HD's. I've unmounted all but the USB and the bad one. @heynnema – john Van Drie Apr 16 '19 at 17:35
  • @johnVanDrie how old is this HDD? It's getting correctable read errors, and it looks like self-test is now failing. Do you have backups of the data? If not, when booted to the Ubuntu Live DVD/USB, can you mount the drive in Files (Nautilus) and copy some stuff off? – heynnema Apr 16 '19 at 17:54
  • @johnVanDrie you ran the fsck when booted to the Ubuntu Live DVD/USB, yes? Did you see my backup comment, above? – heynnema Apr 16 '19 at 18:02
  • Just saw it now. I did a backup to an external drive a while ago (last month?). I'll try Nautilus to get the data. Right now, tho, e2fsck is running - been running 10 min? These HD's are all around 5-7 years old. – john Van Drie Apr 16 '19 at 18:13
  • I probably did something stupid. When I first launch fsck -f and it told me there was a bad magic # in the super-block, it said fsck.ext2: Superblock invalid, trying backup bloks... Superblock needs_recovery flag is clear, but journal has data. Recovery flag not set in backup superblock, so running journal anyway. /dev/sdf1: recovering journal Error reading block 34650768 (Input/output error). Ignore error? yes Force rewrite? cancelled! I'd ctrl-C'd out of this, not knowing what I was doing. JNonethelesss, it ended saying /dev/sdf1: ***** FILE SYSTEM WAS MODIFIED ***** – john Van Drie Apr 16 '19 at 18:15
  • @johnVanDriek if you've got recent enough backups, we can just init the drive by laying down a new partition table, creating a new partition, and running a bad block scan. It's probably just bad. I can document that if you'd like. If you have spare disks, I'd install a fresh Ubuntu there. Show me a gparted screenshot. Let me know. – heynnema Apr 16 '19 at 18:25
  • OK - thanks. The fsck completed, and now I can at least pull off my data from that disk. Will try rebooting once I've done that. And, then, I have a new SSD that I've been meaning to install, I'll do that, and use that as a boot drive. Once I've pulled the key data from my user directory, I think I'll be happy just to ignore this troublesome drive - I have 3 others, and I need to replace them all with modern SSD's. – john Van Drie Apr 16 '19 at 18:37
  • Done! I'm a newbie here, and don't know the ropes. Your help was greatly appreciated - esp. as it was done in real-time. – john Van Drie Apr 16 '19 at 21:44