metal3 - 0.9.0 unexpected pod restart

[hardys]

Since updating the Metal3 chart to 0.9.0 we can see an unexpected pod restart e.g:

metal3-system                     metal3-metal3-ironic-74794b5546-4c4kf                    3/3     Running     1 (3m49s ago)   6m34s

The ironic pod seems to work fine after the restart so this isn't a blocker issue, but we should investigate to understand the root cause - I think it may be caused by the ironic-log-watch container (possibly related to the changes around volumes):

kubectl logs -n metal3-system metal3-metal3-ironic-74794b5546-wrl48 -c ironic-log-watch -p
Setting up watches.
Couldn't watch /shared/log/ironic/deploy: No such file or directory

[hardys]

As discussed in https://gitlab.com/sylva-projects/sylva-core/-/merge_requests/3405