On 6/2/07, James Porter <jameslporter@xxxxxxxxx> wrote:
Glad you are interested in gluster, I absolutely love it! I'm kinda sleepy so if some of my statements are not coherent don't worry about it. Feel free to ask questions. There are several profiles available that let you determine how the gluster bricks will behave. A brick is basically a cluster to the client, you can have multiple bricks. You could set one cluster(all machines) to make your reads fast for your mail daemon to read from. Then you can also configure it to make sure at least 1 copy is on 2 nodes. Instead of a backup server you would have two hot swap servers. You can pretty much configure glusterfs to behave the best for your situation. I personally don't like hardware sitting around doing nothing :) A good place to start: http://gluster.org/docs/index.php/Getting_Started_with_GlusterFS http://gluster.org/docs/index.php/Image:Glusterfs-cluster.png Just try it out, it sells itself. Jim On 6/2/07, Brandon Lamb <brandonlamb@xxxxxxxxx> wrote: > > I am a sysadmin for an ISP and we currently have a single server 20 > scsi drive raid that stores our maildir format mail data, we have > about 105 gigs of mail. > > Our problem is that we have a single point of failure. We have backups > sure but if we lost our drive thats over 5 hours to copy from a backup > and getting a new server up. > > So my question is, can glusterfs help in this area? > > If i were to setup a 3 server cluster and say each server had its own > 200 gig raid, would glusterfs put an identical copy of my mail data on > all 3 servers letting me lose up to 2 servers before i lose my data? > > Or do I completely not understand what glusterfs does? > > I was looking at starfish which sounds similiar to me and they claim > to be a turbo charged google filesystem. > > I was initially looking at a DRBD / OCFS2 solution, but ditched that > in favor of just rebuilding our current box as it is but with a fresh > updated os but remain a single server with no live spare. The decision > was based on keeping it simple and due to not wanting to buy another > scsi raid ($$$). > > But if we could put together 2 servers with SATA II raids (cheap $) > and have glusterfs replicate the data between all 3 servers that would > be cool. > > I look forward to any info this list may provide. > > Thanks! > =) > > Brandon > > > _______________________________________________ > Gluster-devel mailing list > Gluster-devel@xxxxxxxxxx > http://lists.nongnu.org/mailman/listinfo/gluster-devel
not sure if my reply only went to one person so im forwarding directly to list for other new users... HOLY OMG CRAP!!! I just setup a 2 machine cluster with glusterfs. omg. I just got done emailing my two bosses with holy F*#k. This is SWEET!! I set up the config so that both machines exported a local /home/export directory, and then both machines connected to both as clients and setup the config to do the AFR so that files were replicated 2 times. WOW... im just sitting in bed on my laptop grinning at the console screen as i create files on each machine and watch it show up on the other and vice versa. Now I just have to get ZFS working and YES! Kickass! Kudos to the developers. I tried the DRBD/OCFS2 and GFS... and wow, this blows them out the door. I am so stoked in case you couldnt tell, and I havent even done anything with it yet haha. I cant wait to mess around with this at the shop on monday and add more nodes.
