Hello, I recently put a very basic Gluster replication together (2 replicas) and mounted the volume on another server using the Gluster client. However, I was wondering if the 3.5 Gluster client knows about the other replicated volume. In other words, if one of the servers goes down, will the Gluster client know about the replicated volume and keep working? I saw this posting http://www.gluster.org/category/failover/ and it seems that was proposed to replace something like RRDNS but it was unclear if that feature made its way into one more the more recent builds. Thanks! Ryan -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20140624/29d5deba/attachment.html>
JuanFra Rodriguez Cardoso
2014-Jun-24 16:28 UTC
[Gluster-users] Gluster Client and high availability
Hi Ryan: You may use the 'backupvolfile-server' option while mounting. According to RHS documentation: *backupvolfile-server=server name - name of the backup volfile server to mount the client. If this option is added while mounting fuse client, when the first volfile server fails, then the server specified in backupvolfile-server option is used as volfile server to mount the client. * Example: # mount -t glusterfs -o backupvolfile-server=volfile_server2,fetch-attempts=2,log-level=WARNING,log-file=/var/log/gluster.log server1:/test-volume /mnt/glusterfs Regards, --- JuanFra Rodr?guez Cardoso 2014-06-24 18:19 GMT+02:00 Ryan Nix <ryan.nix at gmail.com>:> Hello, > > I recently put a very basic Gluster replication together (2 replicas) and > mounted the volume on another server using the Gluster client. However, I > was wondering if the 3.5 Gluster client knows about the other replicated > volume. In other words, if one of the servers goes down, will the Gluster > client know about the replicated volume and keep working? > > I saw this posting http://www.gluster.org/category/failover/ and it seems > that was proposed to replace something like RRDNS but it was unclear if > that feature made its way into one more the more recent builds. > > Thanks! > > Ryan > > _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > http://supercolony.gluster.org/mailman/listinfo/gluster-users >-------------- next part -------------- An HTML attachment was scrubbed... URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20140624/d173d69d/attachment.html>
Tiziano Müller
2014-Jun-24 16:40 UTC
[Gluster-users] Gluster Client and high availability
Hi Ryan The documentation for GlusterFS is somehow scattered. A good explanation for me was this one: http://joejulian.name/blog/glusterfs-replication-dos-and-donts/ As far as I've understood it (others please correct me if I'm wrong): when a client connects to any server in your GlusterFS cluster and requests access to a volume, he gets the list of servers for that volume and the configuration of the volume. In normal operation it is then the client who does the replication. In case one server crashes, the client still knows the other servers and continues operation using the servers left. As JuanFra already wrote there is an option to specify an alternative server for the initial connection (for getting the configuration), but this has nothing to do with the failover/replication (this is defined while creating the volume), only when mounting the volume. Best regards, Tiziano Am 24.06.2014 18:19, schrieb Ryan Nix:> Hello, > > I recently put a very basic Gluster replication together (2 replicas) and > mounted the volume on another server using the Gluster client. However, I was > wondering if the 3.5 Gluster client knows about the other replicated volume. In > other words, if one of the servers goes down, will the Gluster client know about > the replicated volume and keep working? > > I saw this posting http://www.gluster.org/category/failover/ and it seems that > was proposed to replace something like RRDNS but it was unclear if that feature > made its way into one more the more recent builds. > > Thanks! > > Ryan > > > _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > http://supercolony.gluster.org/mailman/listinfo/gluster-users >-- stepping stone GmbH Neufeldstrasse 9 CH-3012 Bern Telefon: +41 31 332 53 63 www.stepping-stone.ch tiziano.mueller@@stepping-stone.ch