[ovirt-users] Ovirt remove network from hosted engine

knarra knarra at redhat.com
Wed Jun 28 05:29:34 UTC 2017


On 06/27/2017 09:49 PM, Abi Askushi wrote:
> Hi all,
>
> Just in case ones needs it, in order to remove the secondary network 
> interface from engine, you can go to:
> Virtual Machines -> Hostedengine -> Network Interfaces -> edit -> 
> unplug it -> confirm -> remove it.
cool. But in your previous mail you did mention that it fails for you 
since the engine is running. Instead of remove you tried unplug here ?
>
> It was simple...
>
>
> On Tue, Jun 27, 2017 at 4:54 PM, Abi Askushi <rightkicktech at gmail.com 
> <mailto:rightkicktech at gmail.com>> wrote:
>
>     Hi Knarra,
>
>     Then I had already enabled NFS on ISO gluster volume.
>     Maybe i had some networking issue then. I need to remove the
>     secondary interface in order to test that again.
>
>
>
>     On Tue, Jun 27, 2017 at 4:25 PM, knarra <knarra at redhat.com
>     <mailto:knarra at redhat.com>> wrote:
>
>         On 06/27/2017 06:34 PM, Abi Askushi wrote:
>>         Hi Knarra,
>>
>>         The ISO domain is of type gluster though I had nfs enabled on
>>         that volume.
>         you need to have nfs enabled on the volume. what i meant is
>         nfs.disable off which means nfs is on.
>
>         For more info please refer to bug
>         https://bugzilla.redhat.com/show_bug.cgi?id=1437799
>         <https://bugzilla.redhat.com/show_bug.cgi?id=1437799>
>>         I will disable the nfs and try. Though in order to try I need
>>         first to remove that second interface from engine.
>>         Is there a way I can remove the secondary storage network
>>         interface from the engine?
>         I am not sure how to do that, but   you may shutdown the vm
>         using the command hosted-engine --vm-shutdown which will power
>         off the vm and try to remove the networks using vdsclient.
>         (not sure if this is right, but suggesting a way)
>>
>>         Thanx
>
>>
>>         On Tue, Jun 27, 2017 at 3:32 PM, knarra <knarra at redhat.com
>>         <mailto:knarra at redhat.com>> wrote:
>>
>>             On 06/27/2017 05:41 PM, Abi Askushi wrote:
>>>             Hi all,
>>>
>>>             When setting up hosted engine setup on top gluster with
>>>             3 nodes, I had gluster configured on a separate network
>>>             interface, as recommended. When I tried later to upload
>>>             ISO from engine to ISO domain, the engine was not able
>>>             to upload it since the VM did not have access to the
>>>             separate storage network. I then added the storage
>>>             network interface to the hosted engine and ISO upload
>>>             succeeded.
>>             May i know what was the volume type created and added as
>>             ISO domain ?
>>
>>             If you plan to use a glusterfs volume below is the
>>             procedure :
>>
>>             1) Create a glusterfs volume.
>>             2) While adding storage domain select Domain Function as
>>             'ISO' and Storage Type as 'glusterfs' .
>>             3) You can either use 'use managed gluster volume' check
>>             box and select the gluster volume which you have created
>>             for storing ISO's or you can type the full path of the
>>             volume.
>>             4) Once this is added please make sure to set the option
>>             nfs.disable off.
>>             5) Now you can go to HE engine and run the command
>>             engine-iso-uploader upload -i <gluster_iso_domain_name>
>>             <iso_file>
>>
>>             Iso gets uploaded successfully.
>>
>>>
>>>             1st question: do I need to add the network interface to
>>>             engine in order to upload ISOs? does there exist any
>>>             alternate way?
>>             AFAIK, this is not required when glusterfs volume is used.
>>
>>             Attached is the screenshot where i have only one network
>>             attached to my HE which is ovirtmgmt.
>>>
>>>             Then I proceeded to configure bonding for the storage
>>>             domain, bonding 2 NICs at each server. When trying to
>>>             set a custom bond of mode=6 (as recommended from
>>>             gluster) I received a warning that mode0, 5 and 6 cannot
>>>             be configured since the interface is used from VMs. I
>>>             also understood that having the storage network assigned
>>>             to VMs makes it a bridge which decreases performance of
>>>             networking. When trying to remove the network interface
>>>             from engine it cannot be done, since the engine is running.
>>>
>>>             2nd question: Is there a way I can remove the secondary
>>>             storage network interface from the engine?
>>>
>>>             Many thanx
>>>
>>>
>>>             _______________________________________________
>>>             Users mailing list
>>>             Users at ovirt.org <mailto:Users at ovirt.org>
>>>             http://lists.ovirt.org/mailman/listinfo/users
>>>             <http://lists.ovirt.org/mailman/listinfo/users>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20170628/81153127/attachment-0001.html>


More information about the Users mailing list