This message was deleted.
# harvester
a
This message was deleted.
a
That should be coming in Harvester v1.2.2 (see https://github.com/harvester/harvester/issues/5056) which is due for release some time in May (see https://github.com/harvester/harvester/wiki/Roadmap).
Note also that Harvester v1.3.0 is due out this month, which will actually include Longhorn v1.6.0.
b
nice, looking forward on testing Harvester v1.3.0
👍 1
I am curious, why doesn’t Harvester release upgrades by applying manifests in Kubernetes? (I am thinking that release cycles would be faster compared to the current method of building a new .ISO image)
a
Manifests aren't the only thing we need to apply to do an upgrade - we might need to install a newer version of k8s itself, or upgrade the base OS, so only applying manifests isn't enough. There's other considerations too like if we have an ISO with all the necessary images packed into it, you won't run into docker hub rate limits that you might hit pulling images directly, and you can use that ISO to do install/upgrade in airgapped systems.
Building a new ISO itself doesn't take very long (tens of minutes? an hour? I forget - depends how many of the various images you have up-to-date already on your build host). Much more effort goes into making sure we've got all the right bits and pieces together for a given release, and doing QA on the whole lot.
(I wasn't around for the original design of the upgrade process, but if you're interested in some more background, see https://github.com/harvester/harvester/blob/master/enhancements/20220413-zero-downtime-upgrade.md)