spyced
Jonathan Ellis
RT @DavidStrauss: Block devices are the wrong place scale and do HA. It's always expensive (NetApp), unreliable (SPOF), or administrativ ...
Obdurodon
Jeff Darcy
@spyced @DavidStrauss Huh? GlusterFS is *less* administratively complex than e.g. Cassandra. *Far* less. Also, block dev != filesystem.
@spyced @DavidStrauss Huh? GlusterFS is *less* administratively complex than e.g. Cassandra. *Far* less. Also, block dev != filesystem.
DavidStrauss
David Strauss
@Obdurodon GlusterFS is good as it gets but can still get funny under split-brain given the file system approach: community.gluster.org/q/what-gluster…
@Obdurodon GlusterFS is good as it gets but can still get funny under split-brain given the file system approach: community.gluster.org/q/what-gluster…
Obdurodon
Jeff Darcy
@spyced @DavidStrauss It might not be the right choice for any particular case, but for reasons other than administrative complexity.
@spyced @DavidStrauss It might not be the right choice for any particular case, but for reasons other than administrative complexity.
DavidStrauss
David Strauss
@Obdurodon #Gluster has also come a long way in admin complexity, but high-latency (geo) replication still requires manual failover.
@Obdurodon #Gluster has also come a long way in admin complexity, but high-latency (geo) replication still requires manual failover.
Obdurodon
Jeff Darcy
@DavidStrauss I don't think the new behavior (in my answer) is markedly weirder than alternatives, or related to being a filesystem.
@DavidStrauss I don't think the new behavior (in my answer) is markedly weirder than alternatives, or related to being a filesystem.
DavidStrauss
David Strauss
@Obdurodon It's related to it being a filesystem because the consistency model doesn't include a natural, guaranteed split-brain resolution.
@Obdurodon It's related to it being a filesystem because the consistency model doesn't include a natural, guaranteed split-brain resolution.
Obdurodon
Jeff Darcy
@DavidStrauss Those "guarantees" have been routinely violated by most other systems too. I'm not sure why you'd single out just one.
@DavidStrauss Those "guarantees" have been routinely violated by most other systems too. I'm not sure why you'd single out just one.
DavidStrauss
David Strauss
@Obdurodon I'm not singling out Gluster. I think elegant split-brain recovery eludes all distributed POSIX/block device systems.
@Obdurodon I'm not singling out Gluster. I think elegant split-brain recovery eludes all distributed POSIX/block device systems.
Obdurodon
Jeff Darcy
@DavidStrauss I'm loving this conversation, but Twitter might not be the right forum. I'll extract into a blog post.
@DavidStrauss I'm loving this conversation, but Twitter might not be the right forum. I'll extract into a blog post.
DavidStrauss
David Strauss
@Obdurodon You mean complex, theoretical distributed systems issues aren't best handled in 140 characters or less? :-)
@Obdurodon You mean complex, theoretical distributed systems issues aren't best handled in 140 characters or less? :-)
Obdurodon
Jeff Darcy
@DavidStrauss Yes, IMO geosync in its current form is tres lame. That's why I still want to do *real* wide-area replication.
@DavidStrauss Yes, IMO geosync in its current form is tres lame. That's why I still want to do *real* wide-area replication.
DavidStrauss
David Strauss
@Obdurodon Top-notch geo replication requires embracing split-brain as a normal operating mode and having guaranteed, predictable recovery.
@Obdurodon Top-notch geo replication requires embracing split-brain as a normal operating mode and having guaranteed, predictable recovery.
Obdurodon
Jeff Darcy
@DavidStrauss Agreed wrt geo-replication, but that still doesn't support your first general statement since not all systems need that.
@DavidStrauss Agreed wrt geo-replication, but that still doesn't support your first general statement since not all systems need that.
DavidStrauss
David Strauss
@Obdurodon Agreed on need for geo-replication, but geo-repl. issues are just an amplified version of issues experienced in any cluster.
@Obdurodon Agreed on need for geo-replication, but geo-repl. issues are just an amplified version of issues experienced in any cluster.