[pve-devel] ipmi_watchdog : new panic_wdt_timeout param in kernel 4.4

[Alexandre DERUMIER]

>>On the other side, all user space processes are stopped, so it does not really 
>>harm? 

yes, I think it's not a problem . (I never had disk corruption when HA restart vm and source server is kernel panic).

But it's just annoying to wait 4min for reboot (and not documented)


----- Mail original -----
De: "dietmar" <dietmar at proxmox.com>
À: "Alexandre Derumier" <aderumier at odiso.com>, "pve-devel" <pve-devel at pve.proxmox.com>
Envoyé: Jeudi 9 Juin 2016 06:21:40
Objet: Re: [pve-devel] ipmi_watchdog : new panic_wdt_timeout param in kernel 4.4

> I have notice since month that a kernel panic with ipmi watchdog, never 
> restart the server before 4-5min. 
> 
> It seem that ipmi_watchdog had an hardcoded value of 255s when kernel panic 
> occur, to have time to trigger kdump. 
> 
> kernel 4.4 introduce a new option : panic_wdt_timeout , to reduce it 
> 
> https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/?id=c7f42c63901b964833eb23a9bda873b799e7f308 

Wow - that is really strange behavior - they simply overwrite user configured 
values on kernel panic. 

On the other side, all user space processes are stopped, so it does not really 
harm?