AW: [DRBD-user] A simple test

Ryan Steele steele at
Wed Jun 13 19:55:36 CEST 2007

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 your input, and to all the folks who've added their two 
cents.  I have successfully built and test 0.7.24, and can manually fail 
over as described on the DRBD website.  I've RAIDed all my disks, 
including the meta-disk, and I'm very happy with the results I've seen 
so far.   I'll move on to integrating it with Heartbeat2 next, but I 
don't anticipate any problems there. 

Again, thanks - this list has been very insightful and helpful.

Best Regards,

Martin Bene wrote:
> Hi Ryan,
>> Thank you for your input.  I am currently fixing up my FAI 
>> configs such that they do not create an MD device out of 
>> two of the unused partitions, as I don't see any need to 
>> have the DRBD meta-disk striped or mirrored (unless someone 
>> has any good reason pointing to the contrary, e.g. 
>> significant performance gains).
> The fact you're using drbd indicates that you're interested in improving
> the reliablity of your system; while a failover to another system is
> fairly fast it still represents downtime, breaks active network
> connections and can cause transactions to be rolled back in a database.
> Obviosuly, failover sould be avoided if possible, so it makes sense to
> make the seperate nodes fault tolerant as well by using raid 1/5/6.
> Now, _IF_ you're using a raid level that provides reliability for your
> data, you'll want to have some other areas covered as well - even if
> there's no valuable data stored there but just to allow your system to
> keep running in case of a disk failure. Two such areas would be:
>  - system swap
>  - drbd metadata
> Bye, Martin
> _______________________________________________
> drbd-user mailing list
> drbd-user at

More information about the drbd-user mailing list