[DRBD-user] Remove network cable causing system reboot

Louis Voo mailtolouis2020-forum at yahoo.com
Tue Aug 16 19:24:22 CEST 2011

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Thanks for the comment, I'll remove that encryption.

I'm not using any cluster at all, just pure XEN+DRBD+LVM and I don't have fencing device either.
The reason I configure allow-two-primaries is only want to use xen live migration when needed. The system will run on Primary/Secondary at all time, will change to Primary/Primary only  when I need to do maintenance. 

This happen only when I remove the cable, if I do ifdown eth1, it won't happen.


>From: Digimer <linux at alteeve.com>
>To: Louis Voo <mailtolouis2020-forum at yahoo.com>
>Cc: DRBD <drbd-user at lists.linbit.com>
>Sent: Tuesday, August 16, 2011 5:55 PM
>Subject: Re: [DRBD-user] Remove network cable causing system reboot
>On 08/16/2011 12:43 PM, Louis Voo wrote:
>> Hi,
>> I've 2 xen server setup using DRBD. The VM is run on the DRBD partition. 
>> Both server running on Centos 5.6 with 2 NIC, 1 NIC is use by DRBD
>> direct connect to another server using cross over cable. One run on
>> Primary mode and another run on Secondary Mode.
>> Everything work fine so far, but we just found that when the VM is
>> running and if we remove the cross over cable, the primary host will get
>> reboot.
>> Is it something to do to DRBD setup?
>Doesn't look like it, as you've not configured fencing. Are you running
>a cluster proper? If so, what? If you do, and if that cluster is using
>the DRBD link for cluster communication then breaking the link could be
>triggering the cluster's fencing.
>A couple of comments:
>>         net {
>>                 allow-two-primaries;                
>>                 cram-hmac-alg sha1;
>>                 shared-secret "123456";
>You're on a dedicated link, so there is no real reason to add the
>overhead of encryption.
>Secondly, whenever you have dual-primary, you *must* have fencing
>configured. In EL5/DRBD 8.3.9 this is done with:
>        disk {
>                fencing         resource-and-stonith;
>        }
>        handlers {
>                outdate-peer    "/sbin/obliterate-peer.sh";
>        }
>Replace that script if you're not using red hat cluster service with a
>script that will call the fence device of your choice.
>E-Mail:              digimer at alteeve.com
>Freenode handle:     digimer
>Papers and Projects: http://alteeve.com
>Node Assassin:      http://nodeassassin.org
>"At what point did we forget that the Space Shuttle was, essentially,
>a program that strapped human beings to an explosion and tried to stab
>through the sky with fire and math?"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20110816/cbc29791/attachment.htm>

More information about the drbd-user mailing list