Mixing compute and storage without (initial) HCI
Some documentation, especially on older RHEV versions seems to indicate that Gluster storage roles and compute server roles in an oVirt cluster or actually exclusive. Yet HCI is all about doing both, which is slightly confusing when you try to overcome HCI issues simply by running the management engine in a "side car", in my case simply a KVM VM running on a host that's not part of the HCI (attempted) cluster. So the HCI wizard failed to deploy the management VM, but left me with a working Gluster. I then went ahead and used a fresh CentOS 7.7 VM to install the hosted engine the the script variant without an appliance. That went ... not too bad initially, it even found the Gluster bricks eventually, installed the three hosts, looked almost ready to go, but something must be wrong with the "ovirtmgmt" network. It looks alright in the network configuration, where it has all roles, but when I try to create a VM, I cannot select it as the standard network for the VM. Any new network I might try to define, won't work either. So apart from these detail problems: Is the prohibition against mising both roles, gluster and (compute) cluster still in place? Because there are also documentation remnants which seem to indicate that a management engine for an existing cluster can actually be converted into a VM managed by the hosted engine e.g. as part of a recovery, repeating later what the HCI setup attempts to automate.
On Wed, Dec 11, 2019 at 7:46 PM <thomas@hoberg.net> wrote:
Some documentation, especially on older RHEV versions seems to indicate that Gluster storage roles and compute server roles in an oVirt cluster or actually exclusive.
Yet HCI is all about doing both, which is slightly confusing when you try to overcome HCI issues simply by running the management engine in a "side car", in my case simply a KVM VM running on a host that's not part of the HCI (attempted) cluster.
So the HCI wizard failed to deploy the management VM, but left me with a working Gluster.
Indeed, reading latest Red Hat docs, it seems that you have two options: a) if you go with RHHI it is "mandatory" to use hosted engine VM https://access.redhat.com/documentation/en-us/red_hat_hyperconverged_infrast... " Red Hat Hyperconverged Infrastructure for Virtualization (RHHI for Virtualization) combines compute, storage, networking, and management capabilities in one deployment. " and https://access.redhat.com/documentation/en-us/red_hat_hyperconverged_infrast... " 3.4. Hosted Engine virtual machine The Hosted Engine virtual machine requires at least the following: 1 dual core CPU (1 quad core or multiple dual core CPUs recommended) 4GB RAM that is not shared with other processes (16GB recommended) 25GB of local, writable disk space (50GB recommended) 1 NIC with at least 1Gbps bandwidth " b) If you go with "standard" Virtualization with the option of having external engine and you want to adopt Gluster Storage, your storage hosts cannot be at the same time also virtualization hosts: https://access.redhat.com/documentation/en-us/red_hat_gluster_storage/3.4/ht... " 6. Install Red Hat Gluster Storage Install the latest version of Red Hat Gluster Storage on new servers, not the virtualization hosts. " It would be nice a new option to have virtualization and storage node functionalities collapsed on the same hosts, but with an external engine Gianluca
participants (2)
-
Gianluca Cecchi -
thomas＠hoberg.net