It's entirely possible that in a year the k3s project will move away from legacy iptables support and it becomes a non-issue.
That said, Fedora Server on node1 (running kernel 6.19) seems to be performing in-line with the Alpine Linux node2 and node3 for control plane.
I'll let it run this way for a bit just to see how it works in the long run with my normal operations. Maybe later I'll add an agent node.
Not sure why, but upgrading a #k3s test node from #AlpineLinux v3.22 to v3.23 causes etcd to shit itself.
Holding the nodes in the cluster at v3.22 for now until I figure out why.
This is why I have a test cluster that does mostly nothing! Better to FAFO in test than in prod. :)
In today's #AdventuresInSelfHosting: Good lord, I made _ONE_ networking change to my NAS and it completely borked the entire networking stack of the machine. It's been four hours of working on this.
After troubleshooting on my own and with the MFG, the only fix is to reinstall the OS. OMG. Lesson learned: stop touching shit if it's working.
I've shut down my Matrix chat server for now. I really haven't used it a lot, and there's all sorts of questions about the security of the secret messaging the service claims to provide.
https://furry.engineer/@soatok/116094123496968392
I haven't _deleted_ the VM, but it's powered off, and monitoring it is disabled.
Days since I mistook the system shutdown button for the VM shutdown button in Proxmox: 0 š
Fortunately, the two nodes on the K3s cluster that I pin to the system restarted normally and the databases recovered fine.
Reminder for later
system advanced update kernel_extra_options="ixgbe.allow_unsupported_sfp=1"
or
midclt call system.advanced.update '{"kernel_extra_options":"ixgbe.allow_unsupported_sfp=1"}'
After bumping #mastodon server OS to #alpinelinux 3.23, the web server and sidekiq failed to run due to some library version getting upgraded and not found by the gem anymore.
I found someone with a similar issue and saw they were able to re-download all the gems to solve their issue. This worked for me, as well.
Saving this note for later.
Chris @ todfox
Aaron 
bundle install --redownload
Iām slowly starting to figure out Kerberos and LDAP integration, LDAP in general and how to setup ACLs in OpenLDAP.
Now if only I could figure out why Apache Directory Studio (just a tool for browsing LDAP) more often than not has issues connecting using encrypted protocols. The native tools work flawlessly.
Time to abandon the Broadcom-owned "Bitnami" Helm chart and pgSQL images for CloudNative-PG to the CNCF's own charts and images.
Initial deployment in the test cluster seems positive, but now I want to test if I can remove the operator and CRDs without destroying the running cluster. That will be key before I try this on the prod clusters.