Le 25/08/2016 à 13:37, InterNetX - Juergen Gotteswinter a écrit :
> Am 24.08.2016 um 17:15 schrieb InterNetX - Juergen Gotteswinter:
>> iSCSI & Ovirt is an awful combination, no matter if multipathed or
>> bonded. its always gambling how long it will work, and when it fails why
>> did it fail.
We are using oVirt + iSCSI, but using bonding mode 1 (active/passive)
and no multipathing.
It is running OK. So far.
>>
>> its supersensitive to latency, and superfast with setting an host to
>> inactive because the engine thinks something is wrong with it. in most
>> cases there was no real reason for.
You are not wrong...
>>
>> we had this in several different hardware combinations, self built
>> filers up on FreeBSD/Illumos & ZFS, Equallogic SAN, Nexenta Filer
>>
>> Been there, done that, wont do again.
>
> Thank you, I take this as a warning. :)
>
> For my testbed I chose to ignore the iSCSI-bond feature and change the
> multipath default to round robin instead.
>
> What kind of storage do you use in production? Fibre channel, gluster,
> ceph, ...?
>>
>> we had this in several different hardware combinations, self built
>> filers up on FreeBSD/Illumos & ZFS, Equallogic SAN, Nexenta Filer
currently, iscsi multipathed with solaris based filer as backend. but
this is already in progress of getting migrated to a different, less
fragile, plattform. ovirt is nice, but too bleeding edge and way to much
acting like a _girly_
Man, you killed me! :) :) :)
--
Nicolas ECARNOT, crying laughing :) :) :)