29.10.2023 00:07, Zakhar Kirpichenko ?????:> I don't think it's worth it for anyone. It's a dead project since > about 9.0, if not earlier.Well, really earlier. Attempt to get better gluster as gluster2 in 4.0 failed...
W Kern
2023-Oct-31 22:25 UTC
[Gluster-users] [EXT] Re: [Glusterusers] State of the gluster project
Well if what you mean by? 'dead project' is there haven't been significant improvements then yes. Maybe given HOW gluster's architecture works there isn't a lot that can be done to re-architect it. If you mean dead project because Gluster is broken then no. At least for its initial feature set it works really well. We have never used the more advanced features nor did we even try GFAPI. Just vanilla replication w/Fuse We started with Gluster 3.x and it worked well and was easy to manage. Recovering from a failure was bummer though due to the need to heal whole VM files, especially on the 1GB network connections of those days. We then migrated to 6.x and got sharding and the arbiter. Both of which made huge improvements to the speed and recovery in our replication environments. Again, we never had gluster issues with 6.x. The problems we did see were bad networks, drives etc and gluster handled those including the fuse mount keeping the images up during a hardware failure. Then it was a matter of swapping out drives, reassigning volumes etc, all of which were pretty straight forward and didn't involve downtime. We are now on 10.1, and have yet to see any issues. Speed seems a little faster than 6.x but that is subjective. We haven't upgraded to beyond that because we have seen people report issues with 10.2/3/4 and it aint' broke so we have a wait and see attitude. We have used other Distributed File Systems and still use MooseFS for archiving which is quite nice and also easy to use, but as was mentioned with BeeGFS, its freemium. To get the important pieces you have to pay up. In the MFS example that means that the free version has a single point of failure with the mfsmaster. Only enterprise has the ability to failover to another mfsmaster.? So its not as resilient as Gluster and we did lose some files during one particular ugly outage (totally our fault, but those files would have survived if on gluster). Gluster is open source and on github. I hope it stays that way. -wk On 10/29/23 3:54 AM, Dmitry Melekhov wrote:> > 29.10.2023 00:07, Zakhar Kirpichenko ?????: >> I don't think it's worth it for anyone. It's a dead project since >> about 9.0, if not earlier. > > Well, really earlier. > > Attempt to get better gluster as gluster2 in 4.0 failed... > > > ________ > > > > Community Meeting Calendar: > > Schedule - > Every 2nd and 4th Tuesday at 14:30 IST / 09:00 UTC > Bridge: https://meet.google.com/cpu-eiue-hvk > Gluster-users mailing list > Gluster-users at gluster.org > https://lists.gluster.org/mailman/listinfo/gluster-users