INFO:
That worked. Every B server shows CHANGED | rc=0 and "files removed", so the old repo lists are cleared. vlim06 is unreachable, as expected. The "not registered" and Python warning lines are normal on these servers.
Something new to look at: a few servers (vlim02, vlim03, vlip03, vlip06) say "Repository rbb-base is listed more than once in the configuration". That means two .repo files on those servers both define a repo called rbb-base. dnf uses only one of them, so it's a warning, not an error. But if the two copies point at different places, the server may not be looking where you expect. To see both copies:
ansible 'rbb-ref-b-*:!rbb-ref-b-vlim06' -i inventories/ref_rbb/hosts_ref_rbb.yml --ask-vault-pass -u dbv6admin -b -m shell \
-a "grep -n -A3 '^\[rbb-base\]' /etc/yum.repos.d/*.repo"
If both copies have the same baseurl, it's harmless and can be tidied up after patching. If they differ, send me the output before Step 4.
Next: run wizard Step 3 with rbb-ref-b-*:!rbb-ref-b-vlim06. With the cache cleared, vlim03's perl and gcc problems should be gone. If the B master is on version 1.4.2, any remaining problem will come with its dnf: lines.
Comments
Post a Comment