[DRBD-user] Questions regarding offsite disaster/recovery (from stacked drbd device)

Johan Verrept Johan.Verrept at able.be
Tue Oct 27 12:41:56 CET 2009

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


> i've successfully implemented a three-node setup with an offsite node
> via drbd stacked-device as explained in the users-guide
> 
> ( http://www.drbd.org/users-guide/s-pacemaker-stacked-resources.html ) 
> 
> the offsite drbd-device status is always 'secondary' because it has to
> stay in sync with the 'primary' (stacked) device managed by pacemaker
> crm.

Congrats.

> afaik 'secondary' devices cannot be mounted/accessed. 

True. Florian has a post about that on his blog.

http://blogs.linbit.com/florian/2007/06/18/why-drbd-wont-let-you-mount-the-secondary/#more-14

> so, how can i access the offsite data, assumed all nodes in my
> ha-cluster failed? 

I am not quite sure what you want. If you fix the problems with the
cluster and restore the config, you make them sync from your offsite
node. 

If your other nodes failed, I also see no problem in making the offsite
node primary until the other nodes are back up (and in sync with the
offsite node).

> is there a possibility to make the data on the offsite node
> permanently available (read only)? 
>   e.g. so i can always 'clone' my production data to my
> test-environment from the state: 
>     cs:Connected   ro:Secondary/Primary ds:UpToDate/UpToDate 

Use drbd on lvm and make an lvm snapshot of the running drbd device and
mount that snapshot read-only. It isn't live data, but it is as recent
as the snapshot. Also seems good for backups.

	J.





More information about the drbd-user mailing list