Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
On Wed, Nov 03, 2010 at 11:21:34AM -0500, J. Ryan Earl wrote: > So I'd recommend RAID 60 (aka 0+6), with the "0" part being in software, and > the "6" being in hardware; hopefully those controllers have nice backed > write caches. Sounds reasonable. I'll try that setup. > Well, good luck. When SuSE marked "OCFS2" production ready in SLES I think > they did a lot of people a injustice. It was not ready, and speaking for > considerable use of it over the years, there are certain stress cases where > it will fail. Whether you'll hit those cases is a different matter. I've not made my mind up on that. But thanks for that warning! Maybe GFS2 is the better choice, then. > It was unrelated to your original line of questioning which was running a > clustered filesystem on-top DRBD. I have no idea what you're trying to do, > your goals are not clear. Sorry for being unclear! The goal is to have a redundant, high-available file server. I'd like both my boxes to share the same data and being able to failover in case one of them dies. DRBD+CLVM seemed to be a most reasonable solution. My question was indeed not in the line of my original question. I was just wondering how to make this storage accessible to other nodes in the network. iSCSI seems to be the preferred solutions. But maybe I should ask these questions on the linux-ha list. Thanks for your answers so far, they were very helpful! (And improved further research on the topic.) Best regards, Manuel