This message was deleted.
# harvester
a
This message was deleted.
q
Being Harvester based on KubeVirt you can also use their CLI to manipulate VMs without needing to have a UI.
q
great, thanks!
in my case, i had a node that was at 90% cpu for some reason (despite having 0 vms assigned to it) rebooting the node eventually fixed it, just had to get into a good spot to do so.
but a cli would be a great backup
q
Yes, the CLI is definitely handy for those situations.
q
side-note, i thing i have either a) an issue with a host having problems, b) a disk issue, or c) a too much / bottlenecking. any pointers on hunting it down? (6-node cluster, dell r620s, with a bunch of ssds (sata2) via mini-sas jbod
q
The platform itself is quite lean regarding resource consumption, so things like bottlenecks are usually related to the workloads. I recommend you start with the Troubleshooting tips in the documentation: https://docs.harvesterhci.io/v1.1/troubleshooting/installation You may see that a tool called supportconfig is mentioned there several times. If you are a paid customer, that's what you should use to send info to our support colleagues,. If you aren't, it's also useful as it's a package with all the relevant logs altogether. Use it and try to look for errors/warnings in the collected logs. If there's any type of hardware failure or misconfiguration, it's quite likely that you'll something there. If it's a bottleneck or something related to performance, it may not be that obvious, but think about what your VMs are doing and try to follow from there. I hope it helps.
q
@quick-sandwich-76600 thanks! i'll check it out.