Well, that's bad. I guess we're going to have to wait a little longer than we expected... not much of a problem though. AFAIU (poorly) RGW needs to update some index(?) for each write operation and since that index is a single file(?) there is nothing OSD can do about it. That's reasonable at OSD side, not so much at RGW mechanics, sorry if i got things wrong. In the long run IMHO, there should be a way to circumvent it besides faster journal or btrfs. I wish i could be of more help...
Thanks for all the help, you've got a great community here!
On Mon, Feb 18, 2013 at 7:15 PM, Sage Weil <sage@xxxxxxxxxxx> wrote:
On Mon, 18 Feb 2013, Yehuda Sadeh wrote:The problem is actually that it's XFS and so the OSD is doing writeahead
> > Here shows osd receiving the request from rgw and almost immediately queuing
> > it. 0.29 secs pass while this operation gets a chance to actually work. It
>
> Yeah, that's the culprit. That's around ~120 leveldb write operations
> per second. Sam, does that number make sense to you?
journaling, and the operation is a read/modify/write. That serializes the
updates with a journal commit. A faster journal will help, as will using
btrfs... :/
sage
erdem agaoglu
_______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
