Issue with hosted engine restore.

Hello, I have to restore Hosted Engine on another host (completely new hardware). Based on this https://www.ovirt.org/documentation/self-hosted/chap-Backing_up_and_Restorin... is not clear for me if vm's will be rebooted during synchronization hosts with engine ? I have 5 hosts + 1 completely fresh. On host1 I have HE and there is no vm's on other 4 (host1-4) there are around 400 vm which can't be rebooted. Host5 for restore HE. Kris

On Mon, Nov 13, 2017 at 11:58 PM, Krzysztof Wajda <vajdovski@gmail.com> wrote:
Hello,
I have to restore Hosted Engine on another host (completely new hardware). Based on this https://www.ovirt.org/documentation/self-hosted/chap-Backing_up_and_Restorin... is not clear for me if vm's will be rebooted during synchronization hosts with engine ?
They should not be rebooted automatically, but you might need to do this yourself, see below.
I have 5 hosts + 1 completely fresh. On host1 I have HE and there is no vm's on other 4 (host1-4) there are around 400 vm which can't be rebooted. Host5 for restore HE.
Please provide more details about your backup/restore flow. What died (storage? hosts? data?), what are you going to restore, how, etc. Which hosts are hosted-engine hosts. Do they have running VMs. We are working on updating the documentation, but it will take some time. For now, you should assume that the safest way is to pass during restore, to engine-backup, '--he-remove-storage-vm' and '--he-remove-hosts'. This will remove from the engine all the hosted-engine hosts and storage. So when you add the hosts back, you'll have to somehow power off the VMs there - the engine will refuse to add them with running VMs. If you do not want to use these options, you should plan carefully and test. See also: https://bugzilla.redhat.com/show_bug.cgi?id=1235200 https://bugzilla.redhat.com/show_bug.cgi?id=1240466 https://bugzilla.redhat.com/show_bug.cgi?id=1441322 Best regards, -- Didi

Hi David, Thanks for reply, I have Ovirt env which is in very bad shape, and as a new employer I have to fix it :). I have 5 hosts in whole env. There is broken HA for hosted engine, so engine can only works on host1. I can't add another host because it shows me that it was deployed with version 3.5 (If I'm not wrong it's fixed in 4.0). Also I can't update/upgrade ovirt-engine because there is 500MB of free space (after cleaning up) without LVM, so I'm afraid that I'll run out off space during update. Because of that I decided to add completely new server and migrate hosted-engine on fixed HE (with LVM) and properly configured HA on new host. Below short summary: Hosted engine: CentOS Linux release 7.2.1511 (Core) ovirt-engine-3.6.7.5-1.el7.centos.noarch Host0: with running hosted Hosted-Engine which I need to update/upgrade CentOS Linux release 7.2.1511 (Core) ovirt-vmconsole-1.0.0-1.el7.centos.noarch ovirt-hosted-engine-ha-1.3.5.1-1.el7.centos.noarch ovirt-host-deploy-1.4.1-1.el7.centos.noarch ovirt-release36-007-1.noarch ovirt-engine-sdk-python-3.6.3.0-1.el7.centos.noarch ovirt-image-uploader-3.6.0-1.el7.centos.noarch ovirt-setup-lib-1.0.1-1.el7.centos.noarch libgovirt-0.3.3-1.el7_2.1.x86_64 ovirt-vmconsole-host-1.0.0-1.el7.centos.noarch ovirt-hosted-engine-setup-1.3.4.0-1.el7.centos.noarch ovirt-engine-appliance-3.6-20160301.1.el7.centos.noarch vdsm-jsonrpc-4.17.23.2-0.el7.centos.noarch vdsm-yajsonrpc-4.17.23.2-0.el7.centos.noarch vdsm-4.17.23.2-0.el7.centos.noarch vdsm-python-4.17.23.2-0.el7.centos.noarch vdsm-infra-4.17.23.2-0.el7.centos.noarch vdsm-hook-vmfex-dev-4.17.23.2-0.el7.centos.noarch vdsm-xmlrpc-4.17.23.2-0.el7.centos.noarch vdsm-cli-4.17.23.2-0.el7.centos.noarch Output from hosted-engine --vm-status from host1 --== Host 1 status ==-- Status up-to-date : True Hostname : dev-ovirtnode0.example.com Host ID : 1 Engine status : {"health": "good", "vm": "up", "detail": "up"} Score : 3400 stopped : False Local maintenance : False crc32 : d7fdf8b6 Host timestamp : 1243846 --== Host 2 status ==-- <- it's garbage because there is not installed and configured HA on host2 Status up-to-date : False Hostname : dev-ovirtnode1.example.com Host ID : 2 Engine status : unknown stale-data Score : 0 stopped : True Local maintenance : False crc32 : fb5f379e Host timestamp : 563 Remaining Hosts1 - 4 are updated and configured in the same way (it was done by me). I had to replace network cards and now there is LACP on 4x10G cards (before was only 1G card). Because there is CentOS 7.4 I decided to install vdsm in version 4.17.43 (from repo) to fix bugs. I aware that 3.6 is only supported with verision 7.2, but I want to update whole env to 3.6.X then to 4.0. then to 4.1 to be up to date. vdsm-jsonrpc-4.17.43-1.el7.centos.noarch vdsm-xmlrpc-4.17.43-1.el7.centos.noarch vdsm-4.17.43-1.el7.centos.noarch vdsm-infra-4.17.43-1.el7.centos.noarch vdsm-yajsonrpc-4.17.43-1.el7.centos.noarch vdsm-cli-4.17.43-1.el7.centos.noarch vdsm-python-4.17.43-1.el7.centos.noarch vdsm-hook-vmfex-dev-4.17.43-1.el7.centos.noarch On those host1-4 I have around 400 vm's for used by developers, and I need to shorten downtime as much as possible (the best option is without downtime, but I'm not sure if it's possible). I decided to restore HE on completely new host because I believe that in my case it's the easiest way to update then upgrade the whole env :) Many thanks for all advises Regards Krzysztof 2017-11-14 8:50 GMT+01:00 Yedidyah Bar David <didi@redhat.com>:
On Mon, Nov 13, 2017 at 11:58 PM, Krzysztof Wajda <vajdovski@gmail.com> wrote:
Hello,
I have to restore Hosted Engine on another host (completely new hardware). Based on this https://www.ovirt.org/documentation/self-hosted/ chap-Backing_up_and_Restoring_an_EL-Based_Self-Hosted_Environment/ is not clear for me if vm's will be rebooted during synchronization hosts with engine ?
They should not be rebooted automatically, but you might need to do this yourself, see below.
I have 5 hosts + 1 completely fresh. On host1 I have HE and there is no
vm's
on other 4 (host1-4) there are around 400 vm which can't be rebooted. Host5 for restore HE.
Please provide more details about your backup/restore flow. What died (storage? hosts? data?), what are you going to restore, how, etc.
Which hosts are hosted-engine hosts. Do they have running VMs.
We are working on updating the documentation, but it will take some time.
For now, you should assume that the safest way is to pass during restore, to engine-backup, '--he-remove-storage-vm' and '--he-remove-hosts'. This will remove from the engine all the hosted-engine hosts and storage. So when you add the hosts back, you'll have to somehow power off the VMs there - the engine will refuse to add them with running VMs. If you do not want to use these options, you should plan carefully and test.
See also:
https://bugzilla.redhat.com/show_bug.cgi?id=1235200 https://bugzilla.redhat.com/show_bug.cgi?id=1240466 https://bugzilla.redhat.com/show_bug.cgi?id=1441322
Best regards, -- Didi
participants (2)
-
Krzysztof Wajda
-
Yedidyah Bar David