[PVE-User] Proxmox VE 4.1 released!

[Alexandre DERUMIER]

>>3x DELL R620 Servers, with full iDRAC 
>>running current pve from the enterprise repository 

Do you have openmanage installed ?
if yes, please read :
https://pve.proxmox.com/wiki/High_Availability_Cluster_4.x 

I have wrote some notes , because by default openmanage change watchdog timeout and action.

----- Mail original -----
De: "Albert Dengg" <albert at fsfe.org>
À: "proxmoxve" <pve-user at pve.proxmox.com>
Envoyé: Vendredi 11 Décembre 2015 19:59:34
Objet: Re: [PVE-User] Proxmox VE 4.1 released!

hi, 
On Fri, Dec 11, 2015 at 12:24:16PM +0100, Martin Maurer wrote: 
> Hi all, 
> 
> we are proud to announce the final release of our Proxmox VE 4.1! 
:) 

i have a slight problem however and i'm wondering if anybody else 
has this problem: 
while i had fencing working fine using the iTCO watchdog in 4.0, in 
4.1 i cannot get pve to fence a node. 

my setup is the following: 

3x DELL R620 Servers, with full iDRAC 
running current pve from the enterprise repository 


when i try to force a node to fence itself by setting all links to 
down using this simple command 
for i in $(seq 0 5); do ip l s eth$i down; done 
to set all links to down, the node will not reboot itself even 
though pvecm status correctly displays that quorum has been lost 

i have tried: 
iTCO_wdt 
ipmi_watchdog 
both with and without nmi_watchdog=0 

the only differnece is that if i use ipmi_watchdog, i get a message 
in the kernel ring buffer every 5 secondds like that: 
Dec 11 18:59:29 pve2 kernel: [ 361.067004] ipmi_si ipmi_si.0: Could not set the global enables: 0xcc. 

fencing was working before (in fact it seems all nodes where fencing 
themselves on the upgrade from 4.0 to 4.1...) 

does somebody have an idea whats going wrong here or to be more 
exact, where to search next? 

thx 

regards, 
albert 

_______________________________________________ 
pve-user mailing list 
pve-user at pve.proxmox.com 
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user