34

I'm setting up a new Ubuntu 18.04 server and have already played through one backup/restore loop. So the entire system has been restored from backup with a fresh RAID and LVM volumes.

Now recently, unattended-upgrades has reported this:

Processing triggers for initramfs-tools (0.130ubuntu3.8) ...
update-initramfs: Generating /boot/initrd.img-4.15.0-66-generic
I: The initramfs will attempt to resume from /dev/dm-1
I: (/dev/mapper/vg0-swap)
I: Set the RESUME variable to override this.

I tried to find out what that actually means but wasn't successful. What is it resuming there? Some process of generating boot files? I know it can take a while do do that. Does it have to do with the RAID? That is long finished with its resync. I don't get it. And what's actually wrong? What does that message want me to do? I'm clueless. If I don't need it, how can I make it go away?

The server did reboot successfully twice after that message so it can't be important. (If that "I" was supposed to mean something like that.)

Here's some commands if that helps:

> lvs
  LV   VG  Attr       LSize   Pool Origin Data%  Meta%  Move Log Cpy%Sync Convert
  root vg0 -wi-ao---- 400.00g
  swap vg0 -wi-ao----   8.00g

> cat /proc/mdstat
Personalities : [raid1] [linear] [multipath] [raid0] [raid6] [raid5] [raid4] [raid10]
md1 : active raid1 nvme1n1p2[0] nvme0n1p2[1]
      499449152 blocks super 1.2 [2/2] [UU]
      bitmap: 1/4 pages [4KB], 65536KB chunk

md0 : active raid1 nvme0n1p1[1] nvme1n1p1[0]
      523264 blocks super 1.2 [2/2] [UU]

unused devices: <none>

I can't find a good command to show something about these "dm-0" and "dm-1" devices that are also related to the RAID I guess.

ygoe
  • 988

3 Answers3

37

It may happen when you install another OS alongside the first one. Means that swap partition has changed UUID and you need to adjust that in fstab.

To check that run sudo blkid, that gives you actual UUIDs of your swaps, swapon -s will show which one is active, then compare its UUID to the one in /etc/fstab.

If different, edit/create the file /etc/initramfs-tools/conf.d/resume with RESUME=UUID=<swap UUID from blkid>. THen run update-initramfs -u -k all and if it doesn't complain, you win.

If you have encryption setup on your machine also make sure sure cryptsetup-initramfs is still installed before updating initramfs sudo apt-get install cryptsetup-initramfs If you accidentally messed this up, here's an answer to help: Set "older" kernel as default grub entry

Artur Meinild
  • 26,018
Denis Pisarev
  • 471
  • 3
    Interesting, but there is no UUID at all in my fstab. – ygoe Nov 23 '19 at 15:39
  • I changed the swap entry in my fstab from /dev/mapper/kubuntu--vg-swap_1 to UUID="5a69d76a-90b4-4908-9eff-10a144b10b4c" but I still get the same message. – Paul Tomblin Aug 31 '20 at 15:20
  • 2
    I ran into this issue after installing a fresh Ubuntu 20.04 with full disk encryption and doing the first apt update and apt upgrade after logging in. During the first cycle, I paid no attention to the initramfs message mentioned in the question. After a reboot, the system couldn’t boot and I didn’t manage to recover the installation. Upon the second cycle, before rebooting, I followed these and only these steps and now everything seems to be in order. Thank you. – kleinfreund Sep 11 '20 at 15:32
  • 1
    Right Deniis. I tried to install Parrot on a usb. It reformated the Ubu 20 swap and changed the UUID. After changing UBU's fstab to the new swap UUID, the swap was active but boot time was quite slower (on a ssd). Running initramfs fixed it: sudo update-initramfs -u -k all – ross minet Oct 11 '20 at 14:56
  • 1
    I'm in Ubuntu 20.04 with encrypted disk. My swap appeared in /etc/fstab as /dev/mapper/ubuntu--vg-swap_1 and /proc/swaps shows /dev/dm-1. There command update-initramfs was giving me this error. Thing is, /dev/ubuntu-vg/swap_1 actually points to /dev/dm-1 we can check with $ ls -l /dev/ubuntu-vg/swap_1. In any case, I edited the file /etc/initramfs-tools/conf.d/resume as suggested and it worked. Thanks! – Daniel Nov 11 '21 at 21:57
  • I got the message 'update-initranfs' not found. – Michael D Jun 18 '22 at 21:43
  • Here they ARE identical, and I still get the message:

    $ swapon -s Filename Type Size Used Priority /dev/sda7 partition 7811068 436228 -2

    /dev/sda7: UUID="745db9a2-89f1-4c7d-8da9-b57828d2725d" TYPE="swap" PARTUUID="2a0fce6c-07"

    `cat /etc/fstab | grep swap

    swap was on /dev/sda7 during installation

    UUID=745db9a2-89f1-4c7d-8da9-b57828d2725d none swap sw 0 0`

    What else could I do, please?

     – udippel Jul 24 '22 at 12:15
  • For completeness, to oneself: here there just was no file /etc/initramfs-tools/conf.d/resume and I am pretty much sure that it wasn't me to remove it. So I added this file, and the messages went away. – udippel Jul 24 '22 at 14:12
11

This might actually be a (not even new) bug as mentioned here. They and others suggest to add RESUME=none to /etc/initramfs-tools/conf.d/resume.

I needed to sudo-create the file /etc/initramfs-tools/conf.d/resume first on my U 19.04 machine. With sudo update-initramfs -u -k all you test and deploy the initramfs image. See details here.

mfg
  • 1,277
  • 1
    sudo update-initramfs -u -k all doesn't create /etc/initramfs-tools/conf.d/resume for me – tatsu Sep 23 '20 at 16:39
  • 1
    That's not what these tools do. I needed to create the file resume (it wasn't there), and update-initramfs does its own job (installing/setting up a special RAM-based boot file system but what happens there exceeds my knowledge). – mfg Sep 23 '20 at 18:31
  • 1
    This and a BIOS update strangely were what I seem to have needed in my latest iteration of "Why in the hell is an Asus X556UB now failing to boot". Thanks for the info on where the file is (the message should really say where to set RESUME and what a value for F-OFF is). 20.04 here. Around middle of the year this started and has been a royal PITA since. – MrMesees Jan 17 '21 at 03:54
1

Thank You! I love to learn, especially in Linux, because I love Linux! My problem was everytime I got a firmware update I would get the resume variable message, for initd.img from 5.4.0-29 all the way to 5.4.0-100, which took over a half hour, probably longer to update my OS, which is 20.04 that I have been running with no other problems for 3 years now, so after reading I tried a series of commands:

First I ran 'sudo blkid'

Second I ran '/etc/fstab'

Found the same swap uuid

Then ran '/etc/initramfs-tools/conf.d/resume with RESUME=UUID=<swap' and set the ID...

Then ran 'update-initramfs -u -k all'

And it started updating all the old firmwares for me . This has been driving me nuts for over a year now. At first I thought my system was borked. Now I am going to reboot and see if all went well. What kicks me in the ass is that a new OS with LTS is set to be released in a few days. But thats not the point...The point is that I "Learned" something new about linux on what to look for in the future thanks for all the contributions for ideas to try! Maybe someone can get use out of this problem that worked for me.

jbolted
  • 11
  • Hello. It looks like a lot of this answer is just not part of an answer, does not really help to bury the answer under a lot of other stuff. As for a new LTS its more then a few days schedule release is April 21. – David Mar 19 '22 at 06:28
  • 2
    Your answer is too wordy and lengthy. Please [edit] it to be more to the point of what you're saying. After reading it a few times, it's still not clear if you found the problem or not. – LifeBoy Mar 21 '22 at 21:03