On 01/25/2014 01:31 AM, Steve Dainard wrote:
Not sure what a good method to bench this would be, but:
An NFS mount point on virt host: [root@ovirt001 iso-store]# dd if=/dev/zero of=test1 bs=4k count=100000 100000+0 records in 100000+0 records out 409600000 bytes (410 MB) copied, 3.95399 s, 104 MB/s
Raw brick performance on gluster server (yes, I know I shouldn't write directly to the brick): [root@gluster1 iso-store]# dd if=/dev/zero of=test bs=4k count=100000 100000+0 records in 100000+0 records out 409600000 bytes (410 MB) copied, 3.06743 s, 134 MB/s
Gluster mount point on gluster server: [root@gluster1 iso-store]# dd if=/dev/zero of=test bs=4k count=100000 100000+0 records in 100000+0 records out 409600000 bytes (410 MB) copied, 19.5766 s, 20.9 MB/s
The storage servers are a bit older, but are both dual socket quad core opterons with 4x 7200rpm drives.
A block size of 4k is quite small so that the context switch overhead involved with fuse would be more perceivable. Would it be possible to increase the block size for dd and test?
I'm in the process of setting up a share from my desktop and I'll see if I can bench between the two systems. Not sure if my ssd will impact the tests, I've heard there isn't an advantage using ssd storage for glusterfs.
Do you have any pointers to this source of information? Typically glusterfs performance for virtualization work loads is bound by the slowest element in the entire stack. Usually storage/disks happen to be the bottleneck and ssd storage does benefit glusterfs. -Vijay
Does anyone have a hardware reference design for glusterfs as a backend for virt? Or is there a benchmark utility?
*Steve Dainard * IT Infrastructure Manager Miovision <http://miovision.com/> | /Rethink Traffic/ 519-513-2407 ex.250 877-646-8476 (toll-free)
*Blog <http://miovision.com/blog> | **LinkedIn <https://www.linkedin.com/company/miovision-technologies> | Twitter <https://twitter.com/miovision> | Facebook <https://www.facebook.com/miovision>* ------------------------------------------------------------------------ Miovision Technologies Inc. | 148 Manitou Drive, Suite 101, Kitchener, ON, Canada | N2C 1L3 This e-mail may contain information that is privileged or confidential. If you are not the intended recipient, please delete the e-mail and any attachments and notify us immediately.
On Thu, Jan 23, 2014 at 7:18 PM, Andrew Cathrow <acathrow@redhat.com <mailto:acathrow@redhat.com>> wrote:
Are we sure that the issue is the guest I/O - what's the raw performance on the host accessing the gluster storage?
------------------------------------------------------------------------
*From: *"Steve Dainard" <sdainard@miovision.com <mailto:sdainard@miovision.com>> *To: *"Itamar Heim" <iheim@redhat.com <mailto:iheim@redhat.com>> *Cc: *"Ronen Hod" <rhod@redhat.com <mailto:rhod@redhat.com>>, "users" <users@ovirt.org <mailto:users@ovirt.org>>, "Sanjay Rao" <srao@redhat.com <mailto:srao@redhat.com>> *Sent: *Thursday, January 23, 2014 4:56:58 PM *Subject: *Re: [Users] Extremely poor disk access speeds in Windows guest
I have two options, virtio and virtio-scsi.
I was using virtio, and have also attempted virtio-scsi on another Windows guest with the same results.
Using the newest drivers, virtio-win-0.1-74.iso.
*Steve Dainard * IT Infrastructure Manager Miovision <http://miovision.com/> | /Rethink Traffic/ 519-513-2407 <tel:519-513-2407> ex.250 877-646-8476 <tel:877-646-8476> (toll-free)
*Blog <http://miovision.com/blog> | **LinkedIn <https://www.linkedin.com/company/miovision-technologies> | Twitter <https://twitter.com/miovision> | Facebook <https://www.facebook.com/miovision>* ------------------------------------------------------------------------ Miovision Technologies Inc. | 148 Manitou Drive, Suite 101, Kitchener, ON, Canada | N2C 1L3 This e-mail may contain information that is privileged or confidential. If you are not the intended recipient, please delete the e-mail and any attachments and notify us immediately.
On Thu, Jan 23, 2014 at 4:24 PM, Itamar Heim <iheim@redhat.com <mailto:iheim@redhat.com>> wrote:
On 01/23/2014 07:46 PM, Steve Dainard wrote:
Backing Storage: Gluster Replica Storage Domain: NFS Ovirt Hosts: CentOS 6.5 Ovirt version: 3.3.2 Network: GigE # of VM's: 3 - two Linux guests are idle, one Windows guest is installing updates.
I've installed a Windows 2008 R2 guest with virtio disk, and all the drivers from the latest virtio iso. I've also installed the spice agent drivers.
Guest disk access is horribly slow, Resource monitor during Windows updates shows Disk peaking at 1MB/sec (scale never increases) and Disk Queue Length Peaking at 5 and looks to be sitting at that level 99% of the time. 113 updates in Windows has been running solidly for about 2.5 hours and is at 89/113 updates complete.
virtio-block or virtio-scsi? which windows guest driver version for that?
I can't say my Linux guests are blisteringly fast, but updating a guest from RHEL 6.3 fresh install to 6.5 took about 25 minutes.
If anyone has any ideas, please let me know - I haven't found any tuning docs for Windows guests that could explain this issue.
Thanks,
*Steve Dainard *
_________________________________________________ Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> http://lists.ovirt.org/__mailman/listinfo/users <http://lists.ovirt.org/mailman/listinfo/users>
_______________________________________________ Users mailing list Users@ovirt.org <mailto:Users@ovirt.org> http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users