[Users] SLOW I/O performance

[Alex Leonhardt]

Additionally, I've just seen this on the console ovirt engine web console :

Storage domain storage_domain_2 experienced a high latency of 6.76029706001
seconds from host HV2. This may cause performance and functional issues.
Please consult your Storage Administrator


Alex



On 21 January 2013 12:38, Alex Leonhardt <alex.tuxx at gmail.com> wrote:

> Hi All,
>
> This is my current setup:
>
>
> HV1 has :
>  storage_domain_1
>  is SPM master
>
> HV2 has :
>  storage_domain_2
>  is normal (not master)
>
>
> HV1 has storage_domain_1 mounted via 127.0.0.1 (network name, but hosts
> entry sends it to loopback)
>
> HV2 has storage_domain_2 mounted via 127.0.0.1 (network name, but hosts
> entry sends it to loopback)
>
>
> All VMs on HV1 have its storage set to storage_domain_1 and all VMs on HV2
> have their storage set to storage_domain_2
>
>
> My problem now is that after I finally created all the disks on HV2 over a
> super slow mgmt network (ovirtmgmt), it's a 100 Mbit only, I'm now trying
> to kickstart all the VMs I created, however, formatting the disk is taking
> for ever ~ 20-30mins for 12 GB, that is roughly how long it took to create
> the disks over the 100Mbit link.
>
> The weirdness really starts with HV2, as all VMs on HV1 with disks on
> storage_domain_1 have "good" I/O throughput, all VMs on HV2 are awfully
> slow in reading/writing to disk.
>
> I've tried some network settings to increase throughput, but those didnt
> help / had no effect at all.
>
> Anyone come across this issue ? Is it something to do with the ovirtmgmt
> interface only being 100Mbit ?
>
> Alex
>
>
>
>
> --
>
> | RHCE | Senior Systems Engineer | www.vcore.co | www.vsearchcloud.com |
>



-- 

| RHCE | Senior Systems Engineer | www.vcore.co | www.vsearchcloud.com |
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20130121/3dc95550/attachment-0001.html>