11.1 Unable to install linkerd

sritchings

hi all, I'm stuck on step one of lab 11.1
the 'linkerd check' command fails at the linkerd-existence section

only a subset of pods have been created

there seems to be an error in the linkerd-init logs

I have built my cluster on AWS with Rocky(CentOS) with the CRI-O container engine all other labs have worked fine (with the odd error fixed using google) until this one. Any ideas?

chrispokorni

Hi @sritchings,

This may be related to Rocky/Centos/RH no longer using iptables out of the box. Even with iptables manually installed there is no guarantee that it will work smoothly with the applications that need it. And linkerd relies on iptables for traffic filtering.

Based on the earlier VM provisioning and cluster bootstrapping demo videos, on both AWS and GCP, the tested guest OS is Ubuntu LTS (16,18, 20, and I would expect the latest 22 to work as well).

Regards,
-Chris

sritchings

Thanks @chrispokorni ! I was a little concerned about compatibility when I went down the Red Hat / CRI-O route, that's just the Linux distro I'm most familiar with. I was going to rebuild the cluster using Ubuntu anyway before practising the exam to get used to Ubuntu/debian specific commands. Maybe the instructions for cluster setup should call out this issue to save people time as they do contain the alternative setup instructions. I've actually found a pretty straight forward procedure for changing nftables out for iptables on RedHats support site, I may as well try it out anyway before I rebuild the cluster, I'll post the results here.

https://access.redhat.com/solutions/3897641

sritchings

Hi all,
After using the above procedure, flushing iptables and restarting the server I still got the same errors, I'm sure there is a way to fix this but at this point I'm chasing my losses so I'm going to rebuild the cluster using ubuntu/docker.