Proxmox cluster is in a state of development ?

[eurowerfr]

Hi,

I have 2x proxmox 8.3 in cluster with a third.

I have to replace the third.

I try to joint from my PVE 8.3 #4 the cluster : failed. No reason !

Then I delete the cluster in the #4, I reboot, the pveproxy is out !

systemctl start pveproxy
Job for pveproxy.service failed because a timeout was exceeded.
See "systemctl status pveproxy.service" and "journalctl -xeu pveproxy.service" for details.


Anything is OK with Proxmox ?

It's not the first time I have problem with cluster management, it's bug on bug ...

It should be so easy , or with very verbose explanation on what is the problem, but not !

Like all this software developed too quickly in a garage, everything is fine as long as everything works well. But as soon as there is a problem, it's real shit !

Proxmox is so good when all is OK, the GUI, the CLI commands, PBS ... but so much problems ...



Jan 25 12:06:10 xxxxxx pvecm[5559]: malformed JSON string, neither tag, array, object, number, string or atom, at character offset 0 (before "(end of string)") at /usr/share/perl5/PVE/Tools.pm line 1073, <GEN0> chunk 1.
Jan 25 12:06:10 xxxxxx pvecm[5559]: interrupted by unexpected signal

it's very explicit about the reason for the problem ... for a .... fresh installed server ... LOL

 

[maxim.webster]

Like all this software developed too quickly in a garage, everything is fine as long as everything works well. But as soon as there is a problem, it's real shit !

Before continueing your rant, please check your hardware first. Unexpected EOF on files that you didn‘t create is often the consequence of bad ram, bad disk, disk full or broken filesystem.

 

[VictorSTS]

I've installed literally hundreds of clusters with zero issues related to clustering. Yes, there's been bugs with corosync, qdevice, etc but they got sorted out eventually. There aren't too many questions about clustering issues in the forum, so I would bet that clustering works correctly.

Clustering works, so in my very humble opinion the issue is somewhere else: hardware, network, the way PVE was installed/updated/customized or even third party scripts/software installed in the PVE host.

You may have found a new bug, of course! Triple check jorunal, logs and task details, just in case.

 

[eurowerfr]

Thanks for reply.

I do the join cluster from CLI (pvecm) with SSH, instead GUI with API, procedure is ok with CLI and not GUI (for my case)

 

[VictorSTS]

I would bet that something isn't properly installed/updated and you should solve it before continuing deployment, as you may find similar behavior with other actions too.