| [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] |
On Tue, 2003-11-04 at 16:00, Cahill, Ben M wrote: > Greg, thanks for your information ... > > > > > Work is ongoing to solve the lock manager SPOF, but not the > > basic drive > > SPOF issue. > > > > The lock SPOF is being solved via use of OpenDLM as an alternate lock > manager for OGFS. Code is checked in; we need to spend some time to > make sure this work well. > > The drive SPOF solution is outside the scope of a filesystem. > > -- Ben -- > Ben, I agree that solving that should not be done inside the ogfs code. OTOH, having ogfs depend on a external shared disk array to achieve disk failure tolerance is not great either. (ie. The cheapest external shared solutions I know of are $15K and up. And I mean way up.) As far as I know, a cluster enabled version of MD is not on the horizon. For a 2-node ogfs cluster, the www.drbd.org/roadmap.html shows that the drbd team is/was discussing supporting gfs in drbd 0.7, which should automatically support ogfs as well. drbd 0.7 is expected to be released by the end of this year, and has pre-release versions available now. Unfortunately, I don't think this feature was actually incorporated in drbd 0.7, but I'm not positive. For the short term, it is very possible the heartbeat/DRBD/iSCSI setup I described could make a fault tolerant implementation of ogfs possible with comodity parts. In addition, it should scale beyond 2 nodes. If it can be made to work, then I think it would greatly increase the usability of ogfs. Greg -- Greg Freemyer ------------------------------------------------------- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ _______________________________________________ Opengfs-users mailing list Opengfs-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/opengfs-users