Service systemd-hibernate-resume fails on virtual machine after upgrade

lpwevers · 12-08-2020, 03:24 AM

Hi,

I recently upgraded an old server of ours from OpenSuSE 13.2 to 15.2. Upgrade went find and all is working as it should. Except for the fact that every time I reboot it (e.g. for a kernel update) my monitoring complains that it has one failed systemd service.

So, looking at the server I have:

Code:

digs218:~ # systemctl --failed
  UNIT                                             LOAD   ACTIVE SUB    DESCRIPTION
● systemd-hibernate-resume@dev-rootvg-swap.service loaded failed failed Resume from hibernation using device /dev/rootvg/swap

And looking at the status of the service I get this:

Code:

digs218:~ # systemctl status systemd-hibernate-resume@dev-rootvg-swap.service
● systemd-hibernate-resume@dev-rootvg-swap.service - Resume from hibernation using device /dev/rootvg/swap
   Loaded: loaded (/usr/lib/systemd/system/systemd-hibernate-resume@.service; static; vendor preset: disabled)
   Active: failed (Result: exit-code) since Tue 2020-12-08 09:57:11 CET; 11min ago
     Docs: man:systemd-hibernate-resume@.service(8)
 Main PID: 384 (code=exited, status=1/FAILURE)

Dec 08 09:57:11 digs218 systemd[1]: Starting Resume from hibernation using device /dev/rootvg/swap...
Dec 08 09:57:11 digs218 systemd-hibernate-resume[384]: Failed to write '254:1' to /sys/power/resume: Permission denied
Dec 08 09:57:11 digs218 systemd[1]: systemd-hibernate-resume@dev-rootvg-swap.service: Main process exited, code=exited, status=1/FAILURE
Dec 08 09:57:11 digs218 systemd[1]: Failed to start Resume from hibernation using device /dev/rootvg/swap.
Dec 08 09:57:11 digs218 systemd[1]: systemd-hibernate-resume@dev-rootvg-swap.service: Unit entered failed state.
Dec 08 09:57:11 digs218 systemd[1]: systemd-hibernate-resume@dev-rootvg-swap.service: Failed with result 'exit-code'.

Indeed, /sys/power/resume does not exist. And if I try the echo command manually as root I also get the permission denied error.

Now I don't care about hibernating this server at all, so I just tried disabling the service, but it always magically re-enables itself again at each reboot. Is there a way I can get rid of this service? Or fix it; either way is fine for me. The only work-around I have at this moment is running 'systemctl reset-failed' after each reboot.

Kind regards,
Louis

Sauerland · 12-09-2020, 12:44 PM

See here, maybe it can help:
https://wiki.archlinux.org/index.php..._and_hibernate

ondoho · 12-09-2020, 01:57 PM

You don't need this on a virtual machine, disable and, if necessary, mask it.

lpwevers · 12-11-2020, 02:24 AM

Thanks for the replies. Agreed, I don't need it on a VM. I tried masking it, but it still tries to start it:

Code:

digs218:~ # systemctl --failed
  UNIT                                             LOAD   ACTIVE SUB    DESCRIPTION
● systemd-hibernate-resume@dev-rootvg-swap.service masked failed failed systemd-hibernate-resume@dev-rootvg-swap.service

But I did find the cause of it. As expected, it's caused by the upgrade. The old OpenSuSE 13.2 system used the "old" kernelspace method, basically echoing "disk" into /sys/power/state. While after the upgrade the "new" userspace method, handled by the commands s2disk and resume is used. But I guess it never removed the links to the old method.

I've now solved this by changing the resume=/dev/... to noresume in /etc/default/grub and regenerating the grub menu. That prevents the system from trying to resume and thus doesn't try to start this service.