[DRBD-user] Re: Suggestion to prevent split brain situation

Bas van Schaik bas at tuxes.nl
Wed Dec 5 01:02:08 CET 2007

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

Florian Haas wrote:
> Bas,
> On Monday 03 December 2007 11:57:11 Bas van Schaik wrote:
>> Florian Haas wrote:
>>>> Thanks for the clear explanations. I think it needs to be said that I'm
>>>> probably using DRBD in a less professional setup than you guys are.
>>>> Setting up redundant network connections is way beyond my possibilities,
>>>> it simply isn't worth the money and effort. The same holds for the
>>>> Heartbeat setup, but maybe I'll take a better look at that later.
>>> I find it doubtful that it's "way beyond your possibilities" to use that
>>> second on-board NIC that your servers presumably have, and a crossover
>>> cable, for dedicated DRBD replication and a second Heartbeat link.
>> They do not have a second on-board NIC, but I'll consider upgrading them.
>>> However if you're not using Heartbeat, or any cluster manager for that
>>> matter, at all, then there isn't much that will keep you from running
>>> into split brain eventually.
>> That is exactly why I made my suggestions the other day. But you have
>> been clear about this: I should use DRBD in a more professional way, or
>> I shouldn't use it at all.
> It's a little harsh the way you put that. :-) This doesn't have much to do 
> with being more or less professional. But please understand that for the 
> problem you mentioned, a solution already exists in the form of dopd. And 
> please also understand that your alternative suggestion of DRBD refusing to 
> become Primary when disconnected is diametrically opposed to DRBD's primary 
> purpose, namely, to provide storage capability in an HA cluster capable of 
> automatic fail over. And DRBD won't get "fixed" to suit your scenario when 
> that "fix" would break the vast majority of DRBD-based clusters out there.
> Hope this makes my point clearer. Apologies if I sounded rude earlier.
To be completely clear (from the bottom of my heart): no offence
intended, no offence taken.

After rereading your e-mails I decided to upgrade my "cluster" (of two
Xen-servers) a bit. There are two network connections as we speak: a
normal LAN connection (also used for the WAN connection), and an extra
dedicated gigabit connection. But I don't think I understand everything
quite well... Am I right when I state that the whole dopd-thing is only
available for HeartBeat v2 and DRBD 0.8? That would be kind of a problem
for me, since I'm stuck on Debian stable using DRBD 0.7 and HeartBeat
And is it correct that DRBD only supports a single IP per machine using
the "address" directive? In that case the redundant link is only useful
for HeartBeat, but not to DRBD?

Please correct me if I'm wrong! If not: is there an alternative solution
for DRBD 0.7 to mark the data "outdated" using HeartBeat? The ultimate
goal is (of course) to minimize the odds that I will run into a split
brain situation one day...

Thanks again for all the feedback!


  -- Bas

More information about the drbd-user mailing list