[GH-ISSUE #229] Feature request: Ability to mark VMs to always run on different compute nodes #153

New issue

Closed

opened 2026-02-27 15:57:32 +03:00 by kerem · 2 comments

kerem commented

2026-02-27 15:57:32 +03:00

Owner

Originally created by @lord-kyron on GitHub (Mar 25, 2019).
Original GitHub issue: https://github.com/retspen/webvirtcloud/issues/229

@catborise - recently I needed to have several VMs running always on different compute nodes to avoid down time if a compute node crash. Can you add such option to check (mark) in the VM options tab. Something like always run on compute node on which it was initially started, or always run on conpute node {compute node to select}.
And in addition to this the option to avoid running on the same compute node as {possibility to choose several VMs which to avoid}. This way when we have a cluster of VMs for example , we will avoid crash of the whole cluster.
In VMWare it is called host affinity rules or VM/Host rules
Is this achievable?

Originally created by @lord-kyron on GitHub (Mar 25, 2019). Original GitHub issue: https://github.com/retspen/webvirtcloud/issues/229 @catborise - recently I needed to have several VMs running always on different compute nodes to avoid down time if a compute node crash. Can you add such option to check (mark) in the VM options tab. Something like always run on compute node on which it was initially started, or always run on conpute node {compute node to select}. And in addition to this the option to avoid running on the same compute node as {possibility to choose several VMs which to avoid}. This way when we have a cluster of VMs for example , we will avoid crash of the whole cluster. In VMWare it is called host affinity rules or VM/Host rules Is this achievable?

kerem closed this issue

2026-02-27 15:57:32 +03:00

kerem commented

2026-02-27 15:57:33 +03:00

Author

Owner

@lord-kyron commented on GitHub (Mar 25, 2019):

@catborise - I imagine something like this functionality from VMware ESXi:

@lord-kyron commented on GitHub (Mar 25, 2019): @catborise - I imagine something like this functionality from VMware ESXi: ![image](https://user-images.githubusercontent.com/39150122/54920550-288c5c80-4f0c-11e9-9551-da914c538c9e.png)

kerem commented

2026-02-27 15:57:33 +03:00

Author

Owner

@catborise commented on GitHub (Mar 26, 2019):

hi,
this feature is related with cluster. There is not a cluster(pacemaker/corosync) mechanism on webvirtcloud. if there is a compute crash. There is not any mechanism to start them on the other nodes.
Cluster operations are risky and problematic operations.
you mention two things: 1. High Availability, 2. seperation rules

do you using pacemaker or alternatives?

you need this because of random start(CLONE_INSTANCE_AUTO_MIGRATE=True)?

@catborise commented on GitHub (Mar 26, 2019): hi, this feature is related with cluster. There is not a cluster(pacemaker/corosync) mechanism on webvirtcloud. if there is a compute crash. There is not any mechanism to start them on the other nodes. Cluster operations are risky and problematic operations. you mention two things: 1. High Availability, 2. seperation rules do you using pacemaker or alternatives? you need this because of random start(CLONE_INSTANCE_AUTO_MIGRATE=True)?

kerem referenced this issue

2026-02-27 15:58:43 +03:00

[PR #153] [MERGED] fix:sorted btween 'dict' and 'dict' #436