https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# harvester
a
This message was deleted.
t
Changing the storage class is for after harvester is installed. install with the defaults and then modify.
g
oki how how do I modify it? I tried “kubectl edit storageclasses.storage.k8s.io” but no dice
t
Use the GUI? or kubectl?
g
kubectl - GUI is “read only”
I can’t delete it either
t
Ah. the easy answer is create a new storageclass with the settings you want.
g
yes got it, but think when installing rancher-vcluster it used that class, but not 100% - I was unable to upgrade cluster when everything was in replica 3. I am looking for maybe 2x1 node cluster or 1x2 node for my office edge
t
If you have a single host I would dial it down to 1 replica.
g
yes of course and 2 replica for dual nodes. That’s why I am trying to figure this out and just hates it when I don’t understand it all 100% 🙂 - well at least my own configuration
t
just create a new storageclass for each. you should be good from there.
g
just annoying I can delete the default longhorn 3 replica
t
they built it all intending at least 3 nodes with 10gb networking.
g
yes but I can change the default replica but not the other one, just weird. anyway thanks
t
Here is another thing to think about. What ever storage class you use for the image is what will be used for the virtual machine.
g
yes all is good when I spin up new VMs, I just use my own storage class. But just don’t like that I have a storage class with replica 3 that is not being used
t
i wouldn’t worry. 😄
g
when I tried the buildin upgrade, it did not work due to degraded storage / some webhook that had a replica of 3
t
huh. My upgrade worked flawlessly
g
yes I will try it again, first time I did not use pxe with custom config, just ISO install, so could be related to that.
t
that shouldn’t matter on an upgrade.
this time save the upgrade log.
g
I had 2 physical nodes and 1 VM as a witness node runnint 1.3.x and it failed
t
Ah 2 node witness. I have a single node. Save the upgrade log to see what it complains about.
g
and you deployed it with pxe and custom config right?
t
I used iso. But the install process doesn’t matter.
g
ok but then you get replica 3 on all stroage or?
t
yes. But all the volumes are degraded since there is only 1 node.
g
yes and I got upgrade errors related to degraded. let me try to install it again then 1.3.x and get the error logs
t
if you can. just install new with 1.4?
g
I can, but want to be sure I can upgrade
71 Views
Open in Slack
PreviousNext
Powered by