Re: [ovirt-users] Move a VM between 2 setups
Export domains are to move VMs between oVirt environments. As already mentioned, you attach the export domain to one data center, export the vm, maint/detach, attach to different data center, attach and import (then maint/detach, etc.). The language "export domains 'between' engines" is a bit confusing... - Original Message - From: "Christophe TREFOIS" To: "Beckman Daniel" , "users" Cc: Sent: Mon, 14 Nov 2016 20:10:16 + Subject: Re: [ovirt-users] Move a VM between 2 setups Hi Daniel, Fantastic information. Thanks so much for this. I was already using export domains inside my current engine, but not sure if it would also work “between” engines so to speak. Have a great evening! Christophe FROM: Beckman, Daniel [mailto:daniel.beck...@ingramcontent.com] SENT: lundi 14 novembre 2016 20:43 TO: Christophe TREFOIS ; users SUBJECT: Re: [ovirt-users] Move a VM between 2 setups Hi Christophe, An “export domain” is made for just this purpose. Create a NFS (version 3) share and make it accessible to the hypervisors for each engine. (It should be a dedicated NFS share, not used for anything else.) As I recall it should be owned by vdsm:vdsm (36:36). In one of the engines (doesn’t matter which), in the web admin page go to Storage and add a new NFS based export domain, using the NFS share you created. Once it’s activated, test it out; try right-clicking on a VM to “export” it. Note that there can only be one engine connected to a given export domain at any one time. When you’re done testing the export domain on the first engine, you need to put it into “maintenance” and ultimately “detach” it. Then go to the other engine, and this time under Storage instead of “new domain” click “import domain” and enter the same NFS share information. It should recognize that you already have an export domain setup under that NFS share. Attach and activate it, and under Storage / / VM Import, try importing the VM you had previously exported. This is covered (sparsely) in the oVirt documentation at https://www.ovirt.org/documentation/admin-guide/administration-guide/ [1], and it’s covered more coherently in the commercial RHV documentation at https://access.redhat.com/documentation/en/red-hat-virtualization/4.0/single/administration-guide#Storage_properties [2]. Best, Daniel FROM: on behalf of Christophe TREFOIS DATE: Monday, November 14, 2016 at 11:14 AM TO: users SUBJECT: [ovirt-users] Move a VM between 2 setups Hi, We have a setup where we want to deploy 2 engines as the network between 2 buildings is unreliable. With 2 engines, we then want to be able to move VMs (one time) from current engine where they are running to new engine in the other building. Is there a recommended workflow for doing this? We have access to shared NFS for this task if required. Thanks for any pointers, Christophe -- DR CHRISTOPHE TREFOIS, DIPL.-ING. Technical Specialist / Post-Doc UNIVERSITÉ DU LUXEMBOURG LUXEMBOURG CENTRE FOR SYSTEMS BIOMEDICINE Campus Belval | House of Biomedicine 6, avenue du Swing L-4367 Belvaux T: +352 46 66 44 6124 F: +352 46 66 44 6949 http://www.uni.lu/lcsb [6] [7] [8] [9] [10] [11] This message is confidential and may contain privileged information. It is intended for the named recipient only. If you receive it in error please notify me and permanently delete the original message and any copies. Links: -- [1] https://www.ovirt.org/documentation/admin-guide/administration-guide/ [2] https://access.redhat.com/documentation/en/red-hat-virtualization/4.0/single/administration-guide#Storage_properties [3] mailto:users-boun...@ovirt.org [4] mailto:christophe.tref...@uni.lu [5] mailto:users@ovirt.org [6] http://www.uni.lu/lcsb [7] https://www.facebook.com/trefex [8] https://twitter.com/Trefex [9] https://plus.google.com/+ChristopheTrefois/ [10] https://www.linkedin.com/in/trefoischristophe [11] http://skype?call ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] I don't need reports anymore. How do I get rid of it? ovirt 3.6.7
I'd like to get rid of the Jasper server from my mgmt host altogether (not a migration, I just don't want it at all). I disable the dwh and reports services (Centos 7), but I think the engine needs to be configured to not try to send anything (at least it looks like that to me from the logs). What are the steps to removing Reports? I'm using oVirt 3.6.7 ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Import and Exporting domains (to export/import across datacenters) no longer working
In the past we had a 3.4 and a 3.5 ovirt. I was able to attach an Export domain to one, export Templates and/or VMs... put in into Maint mode, Detach it from the datacenter, go to the other ovirt manager, import the Domain (sometimes forcing an attach, sometimes it would automatically attach) and import my VM and/or template. Afterwards, placing it back into Maint mode and Dettaching. Often times there would be a residual icon left under Storage for the detached item... and the process was fairly repeatable. Now, it's getting repeatable in a bad way. Now it won't import, I have to manually do two table record deletes and a REST call to remove the connected storage and then and only then can I successfully import. psql db mods look like (from the ovirt manage box): First to find erroneous data: select * from storage_domain_static where storage_domain_type = 3; Then we locate the offending uuid (see full one below on the REST call) and delete the record our of storage_domain_dynamic and static like so: engine=# delete from storage_domain_dynamic where id = '9f00c1d9-3f2a-41b9-80c3-344900622b07'; DELETE 1 engine=# select Deletestorage_domain_static('9f00c1d9-3f2a-41b9-80c3-344900622b07'); Folks here talked about the latter command, but it would fail if the reference in the storage_domain_dynamic table wasn't removed first. Rest call looks something like this: curl -v -u "admin@internal:ourpassword" -X DELETE https://ovirt.example.com/ovirt-engine/api/storageconnections/9f00c1d9-3f2a-41b9-80c3-344900622b07 Any ideas about what is not apparently broken on our site? Is there any other way to do export/import across datacenters? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Need to clear domain Export already exists in an ovirt (3.5)
Our ovirt 3.5 host thinks it has an export domain, but it's not visible anywhere, but it's keeping us from importing a domain from a different datacenter. What database update do I need to issue to clear the bad state from the ovirt 3.5 we are trying to Import into? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Re: [ovirt-users] Any way to correlate a VM disk (e.g. /dev/vda) to the vdsm ovirt disk?
> Hi ccox, > you can see the disk id mapping to device if you execute 'ls -l > /dev/disk/by-id/' . > Second way, and easier, is to make sure you have guest-agent installed on > your guest virtual machine and using rest API you can run GET command: > GET on .../api/vms/{vm_id}/disks > > You will see an attribute called "" . > I hope that helps should have said I'm running 3.4. I don't think there's a logical_name in that version. And by-id or by-uuid doesn't seem to match anything. Maybe this can't be done in 3.4? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Any way to correlate a VM disk (e.g. /dev/vda) to the vdsm ovirt disk?
I want to correlate virtual disks back to their originating storage under ovirt. Is there any way to do this? e.g. (made up example) /dev/vda maps to ovirt disk disk1_vm serial 978e00a3-b4c9-4962-bc4f-ffc9267acdd8 ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
[ovirt-users] Is a dedicated oVirt mgmt VLAN still needed for oVirt host nodes?
We have an oVirt environment that I inherited. One is running 3.4.0 and one is running 3.5.0. Seem in both cases the prior administrator stated that a dedicated VLAN was necessary for oVirt mgmt. That is, we could not run multiple tagged VLANs on a nic for a given oVirt host node. Does any of this make sense? Is this true? Is it still true for more contemporary versions of oVirt? My problem is that our nodes are blades and I only have two physical nics per blade. In our network for redundancy we need to have the two nics have the same VLANs so that things failover ok. Which means we have to share the oVirt mgmt network on the same wire. That's the ideal. Currently we have a whole nic on the blade just for oVirt management. Is this a requirement? ___ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users