Proxmox server reboots on network drop and VM ssh sessions freeze intermittently

[jneundorf]

HI,

I've got two networking issues with the system that never occurred when I was running VBox and I'd like some help resolving them if possible.

Issue 1: Whenever the my router is rebooted (my switch remains up and running) the proxmox server reboots. In other words - any interruption to the internet will cause the proxmox server to reboot.

Issue 2: ssh client sessions to proxmox vms will freeze from time to time. I have to kill the session and reconnect. It appears to be random.

This is the same hardware I had previous run VBox vms on without the above issues.

Any help would be appreciated.

Thanks,
Jim.

 

[Moayad]

Hello,

Do you enable and use HA?

 

[jneundorf]

Moayad,

Thank you for the information - no I do not, but will investigate.

Thank you,
Jim.

 

[jneundorf]

Moayad,

After looking at the requirements, I only have two nodes so it is not an option.

Thanks for the suggestion anyway.

Jim.

 

[Moayad]

Well, can you please provide us witht he syslog when the issue happened, that will give us more information. You can generate syslog with specific time by issue like the following command:

journalctl --since "2023-08-17 05:30" --until "2023-08-17 09:00" > /tmp/Syslog.txt

You may edit the time/date in the above command.

Thanks for the suggestion anyway.

Sorry that is not a suggested, I asked you if you enabled the HA or not.

 

[jneundorf]

Thanks for the tips Moayad,

I'll post some syslog data soon.

 

[jneundorf]

Moayad,

I rebooted the router and the pve server rebooted approx 14:15. Attached is the syslog from 14:14 to 14:18 FYI.

I appreciate the help.
Jim.

 

[Moayad]

Could you re-generate the syslog entries from 30 minutes before the reboot up to 10 minutes after the reboot?

 

[jneundorf]

Moayad,

Please see attached.

 

[Moayad]

Thank you for the syslog!

Aug 18 13:51:21 pve1 corosync[4699]:   [KNET  ] link: host: 1 link: 0 is down
Aug 18 13:51:21 pve1 corosync[4699]:   [KNET  ] host: host: 1 (passive) best link: 0 (pri: 1)
Aug 18 13:51:21 pve1 corosync[4699]:   [KNET  ] host: host: 1 has no active links
Aug 18 13:51:23 pve1 corosync[4699]:   [KNET  ] rx: host: 1 link: 0 is up
Aug 18 13:51:23 pve1 corosync[4699]:   [KNET  ] link: Resetting MTU for link 0 because host 1 joined
Aug 18 13:51:23 pve1 corosync[4699]:   [KNET  ] host: host: 1 (passive) best link: 0 (pri: 1)
Aug 18 13:51:23 pve1 corosync[4699]:   [KNET  ] pmtud: Global data MTU changed to: 1397

From the provided syslog it seems that you have issue with the network? Can you please provide us with the network configuration and the corosync configuration?

cat /etc/network/interfaces
cat /etc/pve/corosync.conf

Another thing I noticed that the temperature of the hard drive is quite high "75". The high temperature might also cause the system reboot.

Aug 18 14:15:01 pve1 smartd[2844]: Device: /dev/sda [SAT], SMART Usage Attribute: 194 Temperature_Celsius changed from 74 to 75

 

[jneundorf]

Moayad,

Please see attached.

And thank you for the heads-up about the disk temp. I am monitoring the machine via snmp but not the disk temp which I will start. I will need to reconfig the ssd mounts to allow better airflow.

 

[Moayad]

You maybe forget the attach.

 

[jneundorf]

Let's try again ...

I had to zip 'em this time.

 

[Moayad]

Thank you for the config!

We recommend adding a separate NIC for Corosync or adding a second ringX_addr to the Corosync config [0], since you only have two nodes I would add `ring1_addr` to the Corosync config as described in the following link.

[0] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_adding_redundant_links_to_an_existing_cluster

 

[jneundorf]

Moayad,

Thank you for the recommendation. I will work through it and report back.