On Wed, May 5, 2021 at 8:18 AM <ling@aliko.com> wrote:
Hello,
I have been trying to upgrade my self-hosted engine from 4.3 to 4.4 but running into issue while performing hosted-engine deploy. Old hypervisor hosts are all running Centos 8 and old ovirt-engine is also running Centos 7.
I created a brand new baremental node running Centos 8, Kernel 4.18.0-240.15.1.el8_3.x86_64 and the following engine versions: ovirt-hosted-engine-setup-2.4.9-1.el8.noarch ovirt-hosted-engine-ha-2.4.6-1.el8.noarch ovirt-engine-appliance-4.4-20210323171213.1.el8.x86_64 python3-ovirt-engine-sdk4-4.4.10-1.el8.x86_64
I have many VLANs in my environment. But on this host, I only have these network devices set up (eth0 is the main network, eth1 for storage): # nmcli con NAME UUID TYPE DEVICE ovirtmgmt 02f64861-d992-4e56-8cec-da1906bac09f bridge ovirtmgmt
Did you create ovirtmgmt manually, or let the deploy script create it?
System eth1 bd9e565f-bdc3-4e43-bbd3-5875b9d7fed7 ethernet eth1 virbr0 78e6875d-70f6-4c89-89dd-180dbb9250b1 bridge virbr0 eth0 743b0e26-aae7-44b8-9215-3754a537e90b ethernet eth0 vnet0 bcfead6d-c5b6-4428-9f89-41589735be02 tun vnet0
When I run hosted-engine --deploy --restore-from-file=backup_050321.bck, it hangs after showing: [ INFO ] TASK [ovirt.ovirt.engine_setup : Copy yum configuration file] [ INFO ] changed: [localhost -> ovirt.safari.apple.com] [ INFO ] TASK [ovirt.ovirt.engine_setup : Set 'best' to false] [ INFO ] changed: [localhost -> ovirt.safari.apple.com] [ INFO ] TASK [ovirt.ovirt.engine_setup : Update all packages]
You can disable this "Update all packages" by passing e.g.: hosted-engine --deploy --ansible-extra-vars=he_offline_deployment=true
virsh shows the VM is in paused state: # virsh list Id Name State ---------------------------------- 1 HostedEngineLocal paused
Any idea what made it become paused? Please check relevant logs (e.g. /var/log/messages, /var/log/libvirt/qemu/*, /var/log/vdsm/*).
I was able to ssh onto the VM until that point.
You mean, until it became paused? Or also after that?
Do I need to set up all the network connection for all the VLANs before running the deploy script?
In principle you should do nothing - just tell the process which NIC to use for ovirtmgmt, and it will configure ovirtmgmt for you. All the other network configuration you can/should do from inside the engine admin ui. If you have other networks marked "required", the deployment process will fail trying to add the host to the engine (because it can't know which NICs/VLANs should be used for which networks) and will pause, waiting for you to remove some lock file, letting you fix stuff manually. See also: https://bugzilla.redhat.com/1893385 .
And how about engine storage domain? I have a new NFS mount ready but it did not ask me about which storage domain to use. Will it ask in later stage?
Yes, storage is handled (also asked about) in a later stage, after the engine is up and the host added to it.. This is a change from <= 4.2, and was done so that we can use the engine (and vdsm) for this, instead of duplicating their functionality in the deploy code. Good luck and best regards, -- Didi