Hi Heiko, some years ago I had to deliver a reliable storage that should be easy to grow in size over time. For that I was in close contact with presto prime who produced a lot of interesting research results accessible to the public. http://www.prestoprime.org/project/public.en.html what was striking me was the general concern of how and when and with which pattern hard drives will fail, and the rebuilding time in case a "big" (i.e. 2TB+) drive fails. (one of the papers at pp was dealing in detail with that) From that background my approach was to build relatively small raid6 bricks (9 * 2 TB + 1 Hot-Spare) and connect them together with a distributed glusterfs. I never experienced any problems with that and felt quite comfortable about it. That was for just a lot of big file data exported via samba. At the same time I used another, mirrored, glusterfs as a storage backend for my VM-images, same there, no problem and much less hazel and headache than drbd and ocfs2 which I run on another system. hth best
Bernhard Bernhard Glomm IT Administration Phone: +49 (30) 86880 134 Fax: +49 (30) 86880 100 Skype: bernhard.glomm.ecologic Ecologic Institut gemeinnützige GmbH | Pfalzburger Str. 43/44 | 10717 Berlin | Germany GF: R. Andreas Kraemer | AG: Charlottenburg HRB 57947 | USt/VAT-IdNr.: DE811963464 Ecologic™ is a Trade Mark (TM) of Ecologic Institut gemeinnützige GmbH On Dec 9, 2013, at 2:18 PM, Heiko Krämer <hkrae...@anynines.de> wrote: > Signed PGP part > Heyho guys, > > I'm running since years glusterfs in a small environment without big > problems. > > Now I'm going to use glusterFS for a bigger cluster but I've some > questions :) > > Environment: > * 4 Servers > * 20 x 2TB HDD, each > * Raidcontroller > * Raid 10 > * 4x bricks => Replicated, Distributed volume > * Gluster 3.4 > > 1) > I'm asking me, if I can delete the raid10 on each server and create > for each HDD a separate brick. > In this case have a volume 80 Bricks so 4 Server x 20 HDD's. Is there > any experience about the write throughput in a production system with > many of bricks like in this case? In addition i'll get double of HDD > capacity. > > 2) > I've heard a talk about glusterFS and out scaling. The main point was > if more bricks are in use, the scale out process will take a long > time. The problem was/is the Hash-Algo. So I'm asking me how is it if > I've one very big brick (Raid10 20TB on each server) or I've much more > bricks, what's faster and is there any issues? > Is there any experiences ? > > 3) > Failover of a HDD is for a raid controller with HotSpare HDD not a big > deal. Glusterfs will rebuild automatically if a brick fails and there > are no data present, this action will perform a lot of network traffic > between the mirror bricks but it will handle it equal as the raid > controller right ? > > > > Thanks and cheers > Heiko > > > > -- > Anynines.com > > Avarteq GmbH > B.Sc. Informatik > Heiko Krämer > CIO > Twitter: @anynines > > ---- > Geschäftsführer: Alexander Faißt, Dipl.-Inf.(FH) Julian Fischer > Handelsregister: AG Saarbrücken HRB 17413, Ust-IdNr.: DE262633168 > Sitz: Saarbrücken > > <hkraemer.vcf>_______________________________________________ > Gluster-users mailing list > Gluster-users@gluster.org > http://supercolony.gluster.org/mailman/listinfo/gluster-users
signature.asc
Description: Message signed with OpenPGP using GPGMail
_______________________________________________ Gluster-users mailing list Gluster-users@gluster.org http://supercolony.gluster.org/mailman/listinfo/gluster-users