[SOLVED] VM not autobooting when backub is scheduled

[mstefan]

Proxmox-Version: 7.1-1

My server is not live already so I boot and shutdown it every day to work on.
I have an active backup schedule at 02:00 every day.

Always when I boot the server and the scheduled backup did not run because the power was off at night,
the backup starts when proxmox is up and running.

After that the task history tells me, that all VMs get startet.
But no one is started. I have to start it manually... bad for LDAP.
I can not find any errors or sth.

When i reboot after the backup is done, all VMs get started automatically.

Someone got an idea, what is going wrong, that the VMs do not start at boot when backup runs first?

I already tried the VM option Delay, but it did not

 

[dcsapak]

hi,

can you post the journal from a boot, so we can check which things happen when?
also the config from the vm (qm config ID) and the backup schedules

 

[mstefan]

Sorry for the late reply.
Here are the files:

jobs.cfg

vzdump: backup-6d087419-ff46
        schedule 02:00
        compress zstd
        enabled 1
        mailnotification failure
        mailto xxx@xxx.de
        mode snapshot
        prune-backups keep-daily=15,keep-monthly=7,keep-weekly=5,keep-yearly=1
        storage vm_backups
        vmid 100

100.conf

agent: 1
balloon: 2048
boot: order=scsi0;ide2;net0
cores: 6
ide2: none,media=cdrom
memory: 16384
meta: creation-qemu=6.1.0,ctime=1639593496
name: Docker-Host
net0: virtio=02:8E:29:62:05:58,bridge=vmbr0
net1: virtio=F6:99:61:7D:21:F2,bridge=vmbr100
numa: 0
onboot: 1
ostype: l26
parent: zwst004
scsi0: local-dataset:vm-100-disk-0,discard=on,size=32G,ssd=1
scsihw: virtio-scsi-pci
smbios1: uuid=c4fa2db7-a187-4522-8beb-5b4f4239f923
sockets: 1
startup: order=1
vmgenid: 559373f5-fda5-4edb-940b-caaf5c71a0be

journal is attached as file.

 

[dcsapak]

thanks but there is not enough in the logs (it only spans around 7 seconds after the boot) maybe wait a little longer to gather the log (a few minutes at least)

 

[mstefan]

Looks like I am to stupid.
This is everything in there.

Newly booted today and the journalctl stops right after Postfix and man-db regeneration.
Like in the attached file.
10 Minutes later the Output is the same.

EDIT:
Looks like folder2ram makes something strange with my logfiles..
I disable it and will check the log tomorrow.

 

[mstefan]

Now i got a full log. (I dont know what folder2ram did there... will not enable it again)

I took a look at the logsfile and I could not find anything preventing my vm from booting.

• Jan 27 18:35:00: backup job starting for vm100 (finished: Jan 27 18:35:41)
• Jan 27 18:35:02 startall job starting (finished: Jan 27 18:35:12 with OK)

When the backup job is not scheduled the vms boots up.
joournal.txt is from a boot with scheduled backup
journal2 is from a reboot without scheduled backup.

 

[dcsapak]

mhmm. can you post the task logs of the 'startall' tasks and the backup task ?

 

[dcsapak]

ok i found it already. sent a fix[0]
basically, the backup started the vm (temporarily) and pve-guests skipped it because it was already started

with my patch, the vms will get started first, and only then starts the scheduler to execute backups

0: https://lists.proxmox.com/pipermail/pve-devel/2022-January/051521.html

 

[mstefan]

Thanks for your Work.

I just edited my /lib/systemd/system/pvescheduler.service.
Will report back if it solves the problem.

 

[mstefan]

It works! Thank you!