[DRBD-user] using drbd on top of ZFS

Noah Mehl noah at tritonlimited.com
Mon Dec 10 18:01:09 CET 2012

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


I'm not sure how you can expect a setup as describe to even work.  As far as I know, DRBD does not run on solaris/open solaris/illumous, as a key component is a linux kernel module.  Separately, I would NOT recommend running ZFS in linux in production.


On Dec 10, 2012, at 9:21 AM, Maurits van de Lande <M.vandeLande at VDL-Fittings.com<mailto:M.vandeLande at VDL-Fittings.com>> wrote:


I’m thinking of using ZFS as a low level block device for drbd.
ZFS --> drbd --> LVM --> iSCSI

The reason is  as  follows:

- I would like to speed up reading and writing by using large read and write caches (ssd based) (ZFS ZIL and L2ARC).
                Currently my storage is 100% SSD based. It would be nice if  a mixed SSD/HD storage system could be build. $$$.
- I like the copy on write feature of ZFS
- I like the ZFS snapshot feature and being able to replicate the data to a  non drbd system (for backup) http://docs.huihoo.com/opensolaris/solaris-zfs-administration-guide/html/ch06.html
- I have tested Flashcache in the past but this did not work reliable in a two node storage cluster setup. For some reason unknown the cached data was not replicated (correctly) to the other drbd node which of course  resulted in corrupted data.

Does anybody have some experience with this kind of setup?
Are there some cons why this setup should not be used?

Best regards,

Maurits van de Lande
drbd-user mailing list
drbd-user at lists.linbit.com<mailto:drbd-user at lists.linbit.com>

Scanned for viruses and content by the Tranet Spam Sentinel service.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20121210/d5ad9511/attachment.htm>

More information about the drbd-user mailing list