It sounds like you just built a 4GB (or 6GB?) RADOS cluster and then tried
to put 4GB of data into it. That won't work; the underlying local
filesystems probably started having trouble with allocation issues as soon
as you go to 2GB free.
-Greg

On Friday, June 28, 2013, 华仔 wrote:

> Hello, I am from China, I hope you can read my poor english as belows .
>
> We are doing a basic test with ceph and cloudstack.
> *experimental environment :*
> 1.   *four* ceph-osds running on* two* nodes(centos6.2),both of them get
> three 1GB phisical disks(we build osds on /dev/sdb and /dev/sdc).
>       so we get 4GB rbd storage space to use.
>       nodes status: cpu---Intel(R) Xeon(R) CPU L5520  @ 2.27GHz
>                             ram memory---48GB
>                             NIC: Speed: 1000Mb/s
> 2.   one ceph-monitor runnin g on another node(ubuntu13.04).
> 3.   one kvm host node(ceph-client) on which several guest vms run.(we use
> one of them  to do the test)
>
> we test on the guest vm, who has two* rbd-based disks*(*10GB rootdisk and
> 20GB datadisk*).
> we* log on to the vm* and test the *disk's write&read performace* as
> belows.
>
> *write speed*:  wget http://remote server ip/2GB.file ,    we get a write
> speed at an average speed of* 6MB/s*.(*far behind expected*)
> (we must get something wrong there, we would appreciate a lot if any help
> comes from you. *we think the problems comes from the kvm emulator, but
> we are not sure,* *can you give us some advice to improve our vm's disk
> performance in the aspect of writing speed*?)
>
> *read speed*: wget http://local server ip/2GB.file -O /dev/null , the
> average read speed is 39.8MB/s, (that seems great.)
>
> ps: on host server we do the rbd read/write testing too, it works perfect:
> 80MB/s (read/write).
>
>
> Best regards.
>
>
>
>
>

-- 
Software Engineer #42 @ http://inktank.com | http://ceph.com
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com

Reply via email to