At shutdown: "A startjob is running"

fperal · October 5, 2023, 5:23pm

I have a similar issue.
Using Opensuse 15. After the last updates (I’m not sure if the updates are from today or not, as I use to send the computer to sleep instead of power of. I rebooted it today because I saw a message to do so) I got a message

A start job is running from /dev/disk/by-uuid/someid   time running/no limit

and the system won’t boot.

I booted with a rescue system (boot installed system did the trick) and saw the uuid belong to the swap partition, I thought it could be damaged so I reset it with mkswap … the the partition got a new uuid.
I changed the uuid in yast bootloader, recreate the bootloader

fernando@andromeda:~> ls -l /dev/disk/by-uuid
total 0
lrwxrwxrwx 1 root root 15 oct  5 19:06 5b000355-3a1a-49f5-8005-f10668008aa7 -> ../../nvme0n1p1
lrwxrwxrwx 1 root root 10 oct  5 19:06 f030c042-1eb6-47e7-81ed-33dba9eb528d -> ../../sda1
fernando@andromeda:~>

sda1 is swap and the other is the root partition (btrfs)


#mkinitrd
#grub2-install /dev/nvme0n1

I boot … the same, but now the uuid in te message has changed

A start job is running from /dev/disk/by-uuid/f030c042-1eb6-47e7-81ed-33dba9eb528d  time running/no limit

But f030c042-1eb6-47e7-81ed-33dba9eb528d is the right uuid of my swap partition … why the error?

Then I reboot and in grub menu I edit the entry and delete the resume=f030c042-1eb6-47e7-81ed-33dba9eb528d

And the system boots fine.

I can’t see what the problem is

hcvv · October 5, 2023, 5:42pm

When not exactly the subject of this topic, you should always create a new topic. Were it only to expose your problem to more people then only those that follow this one.

Also this is a Tumbleweed case and you say you use an unspecified “Opensuse 15”.

I will split it off.

hcvv · October 5, 2023, 5:45pm

Done.

When the title does not suit you, please ask here for another one. And when LEAP 15.5 is not the openSUSE version you use also mention that.

fperal · October 5, 2023, 6:12pm

It was a typo. I wrote 15. when I thought I had written 15.4

kasi042 · October 5, 2023, 8:24pm

Once I have faced a “startjob is running…” boot failure the related drive wasn’t found as expected from the fstab entry (it was actually unplugged).
Maybe a
cat /etc/fstab
can shed some light on the issue? Probably add a
lsblk -f ?
Does the entry in /dev/disk/by-uuid/THEuuid
point to
/dev/sda1
and does this actually exist?

fperal · October 5, 2023, 9:57pm

It seems there is no problem with the partition.
It is in fstab

UUID=f030c042-1eb6-47e7-81ed-33dba9eb528d   swap         swap   sw,pri=1              0  0

and once I boot edeleting the resume=UUID=f030c042-1eb6-47e7-81ed-33dba9eb528d in grub it gives no problem and in fact the system is using it as swap

fernando@andromeda:~> free -h
              total        used        free      shared  buff/cache   available
Mem:           14Gi       5,4Gi       1,6Gi        65Mi       8,4Gi       9,6Gi
Swap:          59Gi       9,0Mi        59Gi
fernando@andromeda:~>

hcvv · October 6, 2023, 7:40am

Of course when it is in /etc/fstab as swap space.

fperal · October 6, 2023, 10:37am

Yes, of course. Te point is that if is mounting the partition as swap with the uuid set in fstab and is working as swap, why is reporting an error when this same uuid is set in grub in resume=

As far as I know the problem with the resume= in grub could be that the partition is not there, for instance if (as is my case) the swap partition is in other disk than the boot partition and I had unplugged this disk, then the system will be looking for the partition will not find it and I will have an error. But if the partition is there and the uuid is correct and is a swap partition, why I have an error?

fperal · October 6, 2023, 10:40am

I can’t see anything strange

fernando@andromeda:~> ls -l /dev/disk/by-uuid
total 0
lrwxrwxrwx 1 root root 15 oct  5 19:06 5b000355-3a1a-49f5-8005-f10668008aa7 -> ../../nvme0n1p1
lrwxrwxrwx 1 root root 10 oct  5 19:06 f030c042-1eb6-47e7-81ed-33dba9eb528d -> ../../sda1
fernando@andromeda:~> lsblk -f /dev/sda1
NAME FSTYPE FSVER LABEL UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
sda1 swap   1           f030c042-1eb6-47e7-81ed-33dba9eb528d                [SWAP]
fernando@andromeda:~> lsblk -f /dev/nvme0n1p1
NAME      FSTYPE FSVER LABEL UUID                                 FSAVAIL FSUSE% MOUNTPOINTS
nvme0n1p1 btrfs              5b000355-3a1a-49f5-8005-f10668008aa7   90,3G    90% /var
                                                                                 /usr/local
                                                                                 /root
                                                                                 /tmp
                                                                                 /home
                                                                                 /opt
                                                                                 /srv
                                                                                 /.snapshots
                                                                                 /
fernando@andromeda:~>

hcvv · October 6, 2023, 10:49am

Yes, I wonder also. Could it be that it is to small to be used for suspend to disk?

BTW, Swap space is not “mounted”, it is switched on for use.

arvidjaar · October 6, 2023, 10:56am

One obvious possible reason - because necessary drivers are missing in initrd. Your root is NVMe and swap on something else. NVMe drivers are present and drivers for “something else” not.

fperal · October 6, 2023, 11:19am

RAM is 16GB, swap is 64GB

fperal · October 6, 2023, 11:27am

The swap drive is a “normal” sata ssd

#lsinitrd /boot/initrd-5.14.21-150400.24.88-default

........
dracut modules:
systemd
systemd-initrd
i18n
drm
plymouth
btrfs
kernel-modules
kernel-modules-extra
resume
rootfs-block
suse-btrfs
suse-xfs
terminfo
udev-rules
dracut-systemd
haveged
ostree
usrmount
base
fs-lib
shutdown
suse
suse-initrd
.........

arvidjaar · October 6, 2023, 11:46am

Those are dracut modules, not kernel drivers. You can check using e.g.

udevadm info --attribute-walk -p /block/sda | grep DRIVERS=

fperal · October 6, 2023, 12:07pm

fernando@andromeda:~> udevadm info --attribute-walk -p /block/sda | grep DRIVERS=
    DRIVERS=="sd"
    DRIVERS==""
    DRIVERS==""
    DRIVERS==""
    DRIVERS=="ahci"
    DRIVERS=="pcieport"
    DRIVERS==""
fernando@andromeda:~>

arvidjaar · October 6, 2023, 1:29pm

Those are kernel drivers (PCIe should be built-in). You can check whether they are present in initrd:

bor@10:~> sudo lsinitrd /boot/initrd-5.14.21-150400.24.88-default | grep -E 'sd_mod|ahci'
[sudo] password for root: 
-rw-r--r--   1 root     root        34069 Sep 22 16:23 lib/modules/5.14.21-150400.24.88-default/kernel/drivers/ata/ahci.ko.zst
-rw-r--r--   1 root     root        41887 Sep 22 16:23 lib/modules/5.14.21-150400.24.88-default/kernel/drivers/ata/libahci.ko.zst
-rw-r--r--   1 root     root        44875 Sep 22 16:23 lib/modules/5.14.21-150400.24.88-default/kernel/drivers/scsi/sd_mod.ko.zst
bor@10:~>

fperal · October 6, 2023, 2:21pm

I seems they are there

andromeda:~ # lsinitrd /boot/initrd-5.14.21-150400.24.88-default | grep -E 'sd_mod|ahci'
-rw-r--r--   1 root     root        34069 Aug 18 12:36 lib/modules/5.14.21-150400.24.88-default/kernel/drivers/ata/ahci.ko.zst
-rw-r--r--   1 root     root        41887 Aug 18 12:36 lib/modules/5.14.21-150400.24.88-default/kernel/drivers/ata/libahci.ko.zst
-rw-r--r--   1 root     root        44875 Aug 18 12:36 lib/modules/5.14.21-150400.24.88-default/kernel/drivers/scsi/sd_mod.ko.zst
andromeda:~ #