[ovirt-users] Any Kind of Storage IO Limitation?

Yaniv Kaul ykaul at redhat.com
Mon Mar 5 11:34:37 UTC 2018


On Mon, Mar 5, 2018 at 1:25 PM, Thomas Fecke <thomas.fecke at eset.de> wrote:

> Hey Yaniv,
>
>
>
> thanks in advance
>
>
>
> ethtool eno2 | grep Speed
>
>
>
> always brings up 10.000. So i tought there is maybe a config File in ovirt
> that change it or something.
>
>
>
> Or will it use the OS information direct?
>

Correct. I assume you have not put any capping on the performance.


>
>
> Im not sure why the VM´s are getting slow. But when I run an Monitor tool
> to check the Bandwith it never excet 1000. Maybe he just don’t need more
> for the Copy Job
>

We'll need more details on the VM configuration. I assume you are using raw
(and not qcow2), using virtio or virtio-SCSI, enabled IO threads, etc.
What are you using the measure the performance?


>
>
> I have no explanation
>
>
>
> fact ist – when I deploy the same Template about 10-15 Times they are
> getting rly slow.
>

We'll need to understand how you deploy them.


>
>
> The RAM and CPU is morgen then 70% Free.
>
>
>
>
>
>
>
> Ill tried to Stress the Storage a bit.
>
>
>
> Opened nload and created some VM´s, created Templates, created Pools and
> and and
>
>
>
> An I was wrong:
>
>
>
> Max: 8.09 GBit/s
>

That's unlikely. What storage is providing you with 8GB/s?
That needs to be a very very very high end storage....
Y.


>
>
> I can life with that. The Performance issues sees to be Gone for now.
> Maybe it was just the 4.2 Upgrade. I have no idea
>
>
>
> But everything seems to work fine. Sorry for wasting your tim e
>
>
>
>
>
>
>
>
>
> *From:* Yaniv Kaul [mailto:ykaul at redhat.com]
> *Sent:* Montag, 5. März 2018 11:08
> *To:* Thomas Fecke <thomas.fecke at eset.de>
> *Cc:* users at ovirt.org
> *Subject:* Re: [ovirt-users] Any Kind of Storage IO Limitation?
>
>
>
>
>
>
>
> On Mon, Mar 5, 2018 at 8:24 AM, Thomas Fecke <thomas.fecke at eset.de> wrote:
>
> Hey Guys,
>
>
>
> I got a kind of strange Question:
>
>
>
> We got some Hypervisors connected to an x86 Storage ( NFS )
>
>
>
> The Machines are connected via 10 Gbit with that Storage. When I try to
> rsync some Files we reach almost the Maximum Bandwidth.
>
>
>
> But, when I copy some VM´s, Templates or do something Stroage related in
> Ovirt I just can reach 1000 M/bit.
>
>
>
> Is there any kind of “Config Limitation”?
>
>
>
> My biggest Problem:
>
>
>
> We work a lot with Templates. When I deploy 10 VM´s based on one Template
> the VM´s are getting very slow and the Storage seems to be the Problem
>
>
>
> Please verify all your hosts, and especially the SPM, are connected with
> 10g (and have negotiated with 10g and not 1g, etc.)
>
> Y.
>
>
>
>
>
> Thanks a lot
>
>
>
>
>
>
>
>
> _______________________________________________
> Users mailing list
> Users at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20180305/77f195d3/attachment.html>


More information about the Users mailing list