PMG 5.1-3 Freezing

[Rumen Vachev]

After last update of PMG crashed (Freezed) near every day. I have only restart possibility. PMG is instolled over VM on the VMWare 1.8.0 Can you help me?

 

[Stoiko Ivanov]

* Do the logs (files in /var/log) or the journal (journalctl -r) give any indication to the reason of the freeze?
* Does the console show something when the system freezes?
* What are the resources allocated to PMG? (RAM/CPU/Diskspace)

 

[Rumen Vachev]

Hi Mr. Ivanov (Здравейте J)

There is part of syslog before and after PMG freezing.

Configuration of PMG server is:

4x CPU

RAM 6GB

HDD: 100Gb

I am sorry, forget to past log.


Dec 12 20:54:34 pmg postfix/postscreen[21555]: PASS NEW [114.239.251.85]:58082

Dec 12 20:54:35 pmg postfix/smtpd[21557]: connect from unknown[114.239.251.85]

Dec 12 20:54:35 pmg pmgpolicy[20195]: SPF says fail

Dec 12 20:54:35 pmg postfix/smtpd[21557]: NOQUEUE: reject: RCPT from unknown[114.239.251.85]: 554 5.7.1 <recruitment @bg-permrep .eu>: Recipient address rejected: Please see...... proto=SMTP helo=<hengxinshipping.com>

Dec 12 20:54:38 pmg postfix/smtpd[21557]: lost connection after RCPT from unknown[114.239.251.85]

Dec 12 20:54:38 pmg postfix/smtpd[21557]: disconnect from unknown[114.239.251.85] helo=1 mail=1 rcpt=0/1 commands=2/3

Dec 13 10:19:42 pmg systemd[1]: Starting Flush Journal to Persistent Storage...

Dec 13 10:19:42 pmg systemd[1]: Started LVM2 metadata daemon.

 

[Stoiko Ivanov]

Hi ( и Здравейте

The hardware resources should be enough for a moderate workload!

Dec 12 20:54:38 pmg postfix/smtpd[21557]: disconnect from unknown[114.239.251.85] helo=1 mail=1 rcpt=0/1 commands=2/3

Dec 13 10:19:42 pmg systemd[1]: Starting Flush Journal to Persistent Storage...

looks like the VM crashed and got restarted (`Starting Flush Journal to Persistent Storage...` is one of the first not kernel related messages after
a fresh boot)
* Where did you take the logs from? - `journalctl -r` should contain the most information
* Did the PMG reboot? `uptime` should tell you since when it's been running
* Do you see anything relevant in the output of `dmesg`?

 

[Andre Scrivener]

I'm having the same problem, the server is virtual and I'm using vmware.

I already installed version 5.1, 5.0, I installed pure debian and I installed pmg, any ideas?

 

[Stoiko Ivanov]

As before - please check your logs (and the hypervisors logs).
Does PMG reboot or does it continue running?

 

[Andre Scrivener]

Hello Stoiko,

The pmg stops responding and I can not connect via ssh or web.

I will still check the hypervisor logs.

 

[Stoiko Ivanov]

Make sure you have activated persistent journal (the directory `/var/log/journal` has to exist, and if it didn't you need to restart journald) - this gives you the chance to read what happened during the freeze.

Do you need to reset the VM in the hypervisor for it to become responsive again?
Do you see any output on the console of PMG during the freeze?

 

[Andre Scrivener]

Exactly, I need to reset her to respond again.

I can not see any message after the freeze.

When parsing the hypervisor logs, it displays the following message: "The CPU has been disabled by the guest operating system. Power off or reset the virtual machine"

In the manufacturer's KB, it reports that these logs are referring to the operating system's kernel panic.

I noticed that it uses the PVE kernel, can I use the debian kernel and use PMG?

Any help for this?

 

[Stoiko Ivanov]

Any output in the console (of the PMG, when viewed from ESX) when this happens?
Without knowing what causes the panic it's hard to find a workaround/solution.

As for running with debian's kernel - I really cannot say how well this works, since we don't test this setup - but it might be worth a check.

 

[Andre Scrivener]

It shows no output on the console. :/

I'm still trying to find some solution at S.O or Hypervisor.

Any solution, I put here.

 

[netbuild]

It shows no output on the console. :/

I'm still trying to find some solution at S.O or Hypervisor.

Any solution, I put here.

same problem as here... no solution had yet been found.

PMG installed in vmware vsphere 6.5.

all patches and firmwares are up2date, but PMG crashes every day. looks like a kernel problem...

 

[Stoiko Ivanov]

Usually when there's a kernel-problem you see some messages on the console - please post them here - otherwise it's hard to help in debugging

 

[Andre Scrivener]

Hey netubild.

try the following, disable the interface lro.


ethtool -K <name interface> lro off


PS: I forgot to report my problem here.

 

[netbuild]

Usually when there's a kernel-problem you see some messages on the console - please post them here - otherwise it's hard to help in debugging

no, here's nothing. kernel does'nt response to icmp, console shows login prompt and does not react anymore, no reaction on strg+alt+delete

 

[netbuild]

Hey netubild.

try the following, disable the interface lro.


ethtool -K <name interface> lro off


PS: I forgot to report my problem here.

hey, thx... i'll try it

EDIT:

Unfortunately, not work for me

root@mail01:~# sudo ethtool -K ens224 lro off
Cannot get device udp-fragmentation-offload settings: Operation not supported
Cannot get device udp-fragmentation-offload settings: Operation not supported

 

[Stoiko Ivanov]

no, here's nothing. kernel does'nt response to icmp, console shows login prompt and does not react anymore, no reaction on strg+alt+delete

* Try configuring remote syslogging for pmg - and enable persistent journalling (mkdir /var/log/journal) - check what happens after the "freeze"

 

[Andre Scrivener]

hey, thx... i'll try it

EDIT:

Unfortunately, not work for me

root@mail01:~# sudo ethtool -K ens224 lro off
Cannot get device udp-fragmentation-offload settings: Operation not supported
Cannot get device udp-fragmentation-offload settings: Operation not supported

One more detail, install the proxmox on top of a pure debian, I'm not using an image that Proxmox makes available. And, I'm using the debian kernel itself..

Try this, that worked for me.

 

[netbuild]

One more detail, install the proxmox on top of a pure debian, I'm not using an image that Proxmox makes available. And, I'm using the debian kernel itself..

Try this, that worked for me.

thx a lot

seen it as the last option of an unsuccessful process.

 

[netbuild]

* Try configuring remote syslogging for pmg - and enable persistent journalling (mkdir /var/log/journal) - check what happens after the "freeze"

thx, i'll try this.