Hi Bernhard, i have nearly the same setup: - Ubuntu 14.04. - 4 nodes (relplicate distributed) - bricks are on lvm on a mdadm raid 10 used bonding (802.3ad/bond-mode 4) before but did not see real benefit in speed for vms. depending on what you do 1 gb might be already enough. nevertheless we bought 10 ge hardware now For the vms i recommend to use libgfapi which will have better performance and the load of your gluster nodes will go down: http://www.gluster.org/community/documentation/index.php/Libgfapi_with_qemu_libvirt I also used these hints: http://www.gluster.org/community/documentation/index.php/Virt-store-usecase#Tunables Whats also important: http://joejulian.name/blog/keeping-your-vms-from-going-read-only-when-encountering-a-ping-timeout-in-glusterfs/ Regards Andr? Am 18.12.2014 um 12:39 schrieb Bernhard Glomm:> hi all, > > I'm looking for some performance tuning suggestions for glusterfs. > > The set is: > - 2 server with several bricks each (ubuntu 14.04), > - each brick is part of a 2way-mirror between them both > - both server are connected with 2 * 1GB NIC > - the servers also mount the gluster volumes (fuse) > - each gluster volume hosts mainly a single large file (10-20GB) (VM-Image), > sometimes two or three of them > > My qustions: > - Which bond mode would you recommend for this set up? > - what are performance parameters/values for "few large files per gluster volume" > - are there some performance parameters regarding the use as vm.image storage > (the vms do some caching by themself?) > > TIA > > Bernhard > > > > _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > http://supercolony.gluster.org/mailman/listinfo/gluster-users >-- Mit freundlichen Gr??en Andr? Bauer MAGIX Software GmbH Andr? Bauer Administrator Postfach 200914 01194 Dresden Tel. Support Deutschland: 0900/1771115 (1,24 Euro/Min.) Tel. Support ?sterreich: 0900/454571 (1,56 Euro/Min.) Tel. Support Schweiz: 0900/454571 (1,50 CHF/Min.) Email: mailto:abauer at magix.net Web: http://www.magix.com Gesch?ftsf?hrer | Managing Directors: Dr. Arnd Schr?der, Michael Keith Amtsgericht | Commercial Register: Berlin Charlottenburg, HRB 127205 The information in this email is intended only for the addressee named above. Access to this email by anyone else is unauthorized. If you are not the intended recipient of this message any disclosure, copying, distribution or any action taken in reliance on it is prohibited and may be unlawful. MAGIX does not warrant that any attachments are free from viruses or other defects and accepts no liability for any losses resulting from infected email transmissions. Please note that any views expressed in this email may be those of the originator and do not necessarily represent the agenda of the company.
thnx for sharing Andr?! so in ubuntu 14.04 we can use libgfapi without recompiling qemu? it works just out of the box (as described in your first link)? Regards Bernhard> Hi Bernhard, > > i have nearly the same setup: > - Ubuntu 14.04. > - 4 nodes (relplicate distributed) > - bricks are on lvm on a mdadm raid 10 > > used bonding (802.3ad/bond-mode 4) before but did not see real benefit > in speed for vms. depending on what you do 1 gb might be already enough. > nevertheless we bought 10 ge hardware now > > For the vms i recommend to use libgfapi which will have better > performance and the load of your gluster nodes will go down: > > http://www.gluster.org/community/documentation/index.php/Libgfapi_with_qemu_libvirt > > I also used these hints: > http://www.gluster.org/community/documentation/index.php/Virt-store-usecase#Tunables > > Whats also important: > http://joejulian.name/blog/keeping-your-vms-from-going-read-only-when-encountering-a-ping-timeout-in-glusterfs/ > > Regards > Andr? > > Am 18.12.2014 um 12:39 schrieb Bernhard Glomm: >> hi all, >> >> I'm looking for some performance tuning suggestions for glusterfs. >> >> The set is: >> - 2 server with several bricks each (ubuntu 14.04), >> - each brick is part of a 2way-mirror between them both >> - both server are connected with 2 * 1GB NIC >> - the servers also mount the gluster volumes (fuse) >> - each gluster volume hosts mainly a single large file (10-20GB) (VM-Image), >> sometimes two or three of them >> >> My qustions: >> - Which bond mode would you recommend for this set up? >> - what are performance parameters/values for "few large files per gluster volume" >> - are there some performance parameters regarding the use as vm.image storage >> (the vms do some caching by themself?) >> >> TIA >> >> Bernhard >> >> >> >> _______________________________________________ >> Gluster-users mailing list >> Gluster-users at gluster.org >> http://supercolony.gluster.org/mailman/listinfo/gluster-users >> > > > -- > Mit freundlichen Gr??en > > Andr? Bauer > > MAGIX Software GmbH > Andr? Bauer > Administrator > Postfach 200914 > 01194 Dresden > > Tel. Support Deutschland: 0900/1771115 (1,24 Euro/Min.) > Tel. Support ?sterreich: 0900/454571 (1,56 Euro/Min.) > Tel. Support Schweiz: 0900/454571 (1,50 CHF/Min.) > > Email: mailto:abauer at magix.net > Web: http://www.magix.com > > Gesch?ftsf?hrer | Managing Directors: Dr. Arnd Schr?der, Michael Keith > Amtsgericht | Commercial Register: Berlin Charlottenburg, HRB 127205 > > The information in this email is intended only for the addressee named > above. Access to this email by anyone else is unauthorized. If you are > not the intended recipient of this message any disclosure, copying, > distribution or any action taken in reliance on it is prohibited and > may be unlawful. > > MAGIX does not warrant that any attachments are free from viruses or > other defects and accepts no liability for any losses resulting from > infected email transmissions. Please note that any views expressed in > this email may be those of the originator and do not necessarily > represent the agenda of the company. > _______________________________________________ > Gluster-users mailing list > Gluster-users at gluster.org > http://supercolony.gluster.org/mailman/listinfo/gluster-users