How can I disable updates?

[proxmoxniggaz]

I am running Proxmox 5.4-3.
I have about 120VMs running at any one time.

However, yesterday, I logged in to find all of my VMs shutdown, and in the tasks, it noted there was an update which was not user triggered by me but automatic.
It was quite a pain in the ass as I had to restart all the VMs.

How can I disable these auto update downloads from terminal/command line?

Please tell me the exact thing I must type as I am not too savvy with this kinda thing - thanks in advance!

 

[tom]

Proxmox VE does not install packages automatically - if you see this, someone (you) enabled this.?

If you talk about the VMs, you need to check what you installed and how you configured it inside the VM.

 

[proxmoxniggaz]

No one enabled it. Happened whilst I was sleeping... Ive several servers and it only occured on this one

 

[oguz]

hi,

However, yesterday, I logged in to find all of my VMs shutdown, and in the tasks, it noted there was an update which was not user triggered by me but automatic.

'update' is done automatically when the node is booted, but it doesn't "upgrade" anything. you can see this in the task log on every boot.

so my guess is that the problem is something else completely, which likely caused your node to crash or reboot...

i suggest you check the logs and journals to find the actual issue. (/var/log/syslog is the first file you should check. use 'cat' or 'less' command)

 

[proxmoxniggaz]

Is there a way to disable updates? Perhaps it may have crashed. Though I did not notice anything of that sort. Just in the taskbar logs something about a failed update (whether it was being autodownloaded or whatever I do not know as I was asleep and am the sole admin).

 

[t.lamprecht]

Updates did not crash anything for sure, it doesn't touch any file related to VMs running.

The failed update could have been just a symptom of your real issue. Please check your syslog/journal, find out when the VMs got stopped, see if something happened there. You say you have 120 VMs, on a single node? Maybe the node got overloaded, e.g., out of memory and the OOM-killer triggered, just as a educated guess.

 

[Phunny]

i have two nodes with that nasty update behaviour, so every few days up to 10 days the machine dies, acording the log
allways if the host dies the last entry of update try and failure with timesteamp matching the power down time.
as this are both DL380 G7 they stay yellow powered of and the error led flashing, this state not even could be resolved with Ilo and have to be taken care onsite.
I prefer to stop the update stuff completely, but i never found a answer to this question how to stop this

 

[leesteken]

If you comment out (or remove, but you'll might need them later) all the lines (by adding # in front of them) in the files in /etc/apt/sources.list.d/ directory and the /etc/apt/sources.list file, the system won't be able to run apt update (which checks for new updates).
Better would be to find out what is running apt upgrade or apt dist-upgrade (which installes updates but does not stop VMs), and stop running that. Or maybe someone installed other software to install updates and stop VMs? I would not know how to stop that.
This is not standard Proxmox behavior, so I don't know where to begin looking. Proxmox by default does not install updates and it does not stop VMs before, during or after updates. Reinstalling Proxmox from scratch to get the default behavior might be an option?

 

[Phunny]

Well proxmox run every 24h an apt-get which fails .. why apt-get fails in a plain fresh installed system is someting else, but after a few retrys the host die. not everytime but when it does then straight after the failed apt-get
and i didnt mean stoped a VM but thew host power off even in the log is mentioned a simple power off
there is nothing additional installed nor has someone other than me access.
on such hardware i run 3 ESXi since long time without a similiar behaviour, swapping the Server didn't help

 

[leesteken]

apt-get update fails if you dont' have a subscription or have not disabled the enterprise repository. You can just remove /etc/apt/sources.list.d/pve-enterprise.list to get rid of this error.
I just found out (by runniong journalctl | grep apt) that there is a systemd service apt-daily-upgrade that runs every day. It does not give my any trouble and it does not install upgrades or shutdown my server, but it also does not fail on the repository because I configured the no-subscription repository.
If you think that apt-daily-upgrade is the cause of the host dying or shutting down, try disabling it running every day with systemctl disable apt-daily-upgrade.timer.
My guest is that someone or something else is triggering a shutdown, which happens to be around the same time. Maybe externally, like a power button press or UPS signal due to power delivery issue? Maybe you can share some journalctl logs from around the time of shutdown?