At this point I'd go to engine VM and remove host from the postgres DB manually. 
A bit of a hack, but...

ssh root@<engine>
su - postgres
cd /opt/rh/rh-postgresql10/
 source enable
psql engine
select vds_id from vds_static where host_name='myhost1.mydomain.com';
select DeleteVds('<vds_id from prev statement>');

Of course, keep in mind that editing database directly is the last resort and not supported in any way. 

--
Dmitry Filonov
Linux Administrator
SBGrid Core | Harvard Medical School
250 Longwood Ave, SGM-114
Boston, MA 02115


On Mon, Jun 10, 2019 at 11:16 AM Adrian Quintero <adrianquintero@gmail.com> wrote:
Ok I have tried reinstalling the server from scratch with a different name and IP address and when trying to add it to cluster I get the following error:

Event details
ID: 505
Time: Jun 10, 2019, 10:00:00 AM
Message: Host myshost2.virt.iad3p installation failed. Host myhost2.mydomain.com reports unique id which already registered for myhost1.mydomain.com

I am at a loss here, I don't have a brand new server to do this and in need to re-use what I have.


From the oVirt engine log (/var/log/ovirt-engine/engine.log):
2019-06-10 10:57:59,950-04 ERROR [org.ovirt.engine.core.dal.dbbroker.auditloghandling.AuditLogDirector] (EE-ManagedThreadFactory-engine-Thread-37744) [9b88055] EVENT_ID: VDS_INSTALL_FAILED(505), Host myhost2.mydomain.com installation failed. Host myhost2.mydomain.com reports unique id which already registered for myhost1.mydomain.com

So in the /var/log/ovirt-engine/host-deploy/ovirt-host-deploy-20190610105759-myhost2.mydomain.com-9b88055.log of the ovirt engine I see that the host deploy is running  the following command to identify the system, if this is the case then it will never work :( because it identifies each host using the system uuid.

dmidecode -s system-uuid
b64d566e-055d-44d4-83a2-d3b83f25412e


Any suggestions?

Thanks

On Sat, Jun 8, 2019 at 11:23 AM Adrian Quintero <adrianquintero@gmail.com> wrote:
Leo,
I did try putting it under maintenance and checking to ignore gluster and it did not work.
Error while executing action: 
-Cannot remove host. Server having gluster volume.

Note: the server was already reinstalled so gluster will never see the volumes or bricks for this server.

I will rename the server to myhost2.mydomain.com and try to replace the bricks hopefully that might work, however it would be good to know that you can re-install from scratch an existing cluster server and put it back to the cluster.

Still doing research hopefully we can find a way.

thanks again

Adrian




On Fri, Jun 7, 2019 at 2:39 AM Leo David <leoalex@gmail.com> wrote:
You will need to remove the storage role from that server first ( not being part of gluster cluster ).
I cannot test this right now on production,  but maybe putting host although its already died under "mantainance" while checking to ignore guster warning will let you remove it.
Maybe I am wrong about the procedure,  can anybody input an advice helping with this situation ?
Cheers,

Leo




On Thu, Jun 6, 2019 at 9:45 PM Adrian Quintero <adrianquintero@gmail.com> wrote:
I tried removing the bad host but running into the following issue , any idea?

Operation Canceled

Error while executing action: 

host1.mydomain.com
  • Cannot remove Host. Server having Gluster volume.



On Thu, Jun 6, 2019 at 11:18 AM Adrian Quintero <adrianquintero@gmail.com> wrote:
Leo, I forgot to mention that I have 1 SSD disk for caching purposes, wondering how that setup should be achieved?

thanks,

Adrian

On Wed, Jun 5, 2019 at 11:25 PM Adrian Quintero <adrianquintero@gmail.com> wrote:
Hi Leo, yes, this helps a lot, this confirms the plan we had in mind.

Will test tomorrow and post the results.

Thanks again

Adrian

On Wed, Jun 5, 2019 at 11:18 PM Leo David <leoalex@gmail.com> wrote:
Hi Adrian,
I think the steps are:
- reinstall the host
- join it to virtualisation cluster
And if was member of gluster cluster as well:
- go to host - storage devices
- create the bricks on the devices - as they are on the other hosts
- go to storage - volumes
- replace each failed brick with the corresponding new one.
Hope it helps.
Cheers,
Leo


On Wed, Jun 5, 2019, 23:09 <adrianquintero@gmail.com> wrote:
Anybody have had to replace a failed host from a 3, 6, or 9 node hyperconverged setup with gluster storage?

One of my hosts is completely dead, I need to do a fresh install using ovirt node iso, can anybody point me to the proper steps?

thanks,
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/RFBYQKWC2KNZVYTYQF5T256UZBCJHK5F/
--
Adrian Quintero


--
Adrian Quintero


--
Adrian Quintero


--
Best regards, Leo David


--
Adrian Quintero


--
Adrian Quintero
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/4XZVAK7S7CVLG72TUJKQFU74P3Q4NVTC/