Hi, I made the test : I have 3 nodes on Centos 6.3 for mon/mds and 8 nodes under Debian Squeeze with 3.2.0 kernel for osd (each with 1 hd with 1To for osd and the journal on the OS hd) and for mounting cephfs (i know it's very very bad... but i use ceph for a ro filesystem shared on all node of a compute cluster... don't see problem in ro access, and performance are excellent ;)).All ceph component are in 0.56.4 $ time tar -zxf Joomla_3.0.3-Stable-Full_Package.tar.gzreal 0m20.398suser 0m0.512ssys 0m0.720s On local hd :$ time tar -zxf Joomla_3.0.3-Stable-Full_Package.tar.gzreal 0m0.356suser 0m0.348ssys 0m0.144s
I don't know if it's a usefull test... use iozone & cie for bench is more adequat in my opinion. From: timo...@koolin.ru Date: Thu, 11 Apr 2013 12:12:31 +0400 To: ceph-users@lists.ceph.com Subject: Re: [ceph-users] Performance of cephfs Can anybody repeat is test in own production cluster? 2013/4/4 Timofey Koolin <timo...@koolin.ru> I have centos 6.3 with kernel 3.8.4-1.el6.elrepo.x86_64 from elrepo.org. Cephfs mount with kernel module. [root@localhost t1]# wget http://joomlacode.org/gf/download/frsrelease/17965/78413/Joomla_3.0.3-Stable-Full_Package.tar.gz [root@localhost t1]# time tar -zxf Joomla_3.0.3-Stable-Full_Package.tar.gz real 8m58.914suser 0m0.414s sys 0m0.757s Same on local file system:[root@localhost t1]# time tar -zxf Joomla_3.0.3-Stable-Full_Package.tar.gz real 0m0.477suser 0m0.252ssys 0m0.229s I have empty cluster with 3 VDS on same host server. Every VDS have 1GB RAM. Have have no any load in this cluster. Is it normal?-- Blog: www.rekby.ru -- Blog: www.rekby.ru _______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
_______________________________________________ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
