Currently, the initContainer do all the "entrypoint" process used by Peertube.
While it allows to speed up (a lot) Peertube start by avoiding the massive chown on all existing videos, it also disable the cp default that init default parameters at start.
But when Peertube is upgraded, if this cp is not done, it may miss some new parameters, which causes Peertube does not start.
So we need to find a way either to execute this cp independently of the chown, or detect new image through helm and execute the cp if a new image is used.
Hello,
Currently, the initContainer do all the "entrypoint" process used by Peertube.
While it allows to speed up (a lot) Peertube start by avoiding the massive chown on all existing videos, it also disable the `cp default` that init default parameters at start.
But when Peertube is upgraded, if this `cp` is not done, it may miss some new parameters, which causes Peertube does not start.
So we need to find a way either to execute this `cp` independently of the chown, or detect new image through helm and execute the cp if a new image is used.
Hello,
Currently, the initContainer do all the "entrypoint" process used by Peertube.
While it allows to speed up (a lot) Peertube start by avoiding the massive chown on all existing videos, it also disable the
cp default
that init default parameters at start.But when Peertube is upgraded, if this
cp
is not done, it may miss some new parameters, which causes Peertube does not start.So we need to find a way either to execute this
cp
independently of the chown, or detect new image through helm and execute the cp if a new image is used.