examine large chown'ed directories in container images and see if they can be reduced

[mmguero]

sparked from #861 and things like podman-container-tools/podman#16541

In the entrypoint script shared by all containers there are some chown operations that have to occur to make sure files/directories are owned by the right UID/GID. When the user running Malcolm (PUID/PGID in process.env) matches what was there at build time (1000/1000 by default) this isn't too expensive. But when the user is different, it can take a long time, particularly on rotational backing media.

Some of this can't be avoided. However, I think that there are some large directories (large as in numbers of files, not file size) that are being blanket-chowned that we could probably be more targeted on.

zeek-live, logstash, dashboards, and netbox seem to be some of the worst offenders.

We ought to:

1. for testing purposes (not committed to GitHub) add timing output to the running of docker-uid-gid-setup.sh
2. determine the ones that take the longest (as a non-1000 uid/gid user)
3. figure out which directories/subdirectories could be excluded from this, if possible

[divinehawk]

Noticed this: zeek:26.02.0 is uid 0 in /usr/local/zeek, so even 1000 gets a costly chown... at least 15 minutes on my system (not SSD).

[mmguero]

I recently changed base images to the docker image, and it could be that I need tweak that. thanks for that feedback.

[mmguero]

I put some debugging on my entrypoint to find out the most obvious offenders:

dashboards-1  | PUSER_CHOWN Duration: 0.001 s
dashboards-1  | find chown Duration: 0.396 s
dashboards-1  | find chmod Duration: 104.176 s
zeek-1        | PUSER_CHOWN Duration: 0.197 s
zeek-1        | find chown Duration: 0.404 s
zeek-1        | find chmod Duration: 0.226 s
netbox-1        | PUSER_CHOWN Duration: 0.002 s
netbox-1        | find chown Duration: 38.462 s
netbox-1        | find chmod Duration: 0.282 s
logstash-1           | PUSER_CHOWN Duration: 0.007 s
logstash-1           | find chown Duration: 0.242 s
logstash-1           | find chmod Duration: 0.181 s
zeek-live-1          | PUSER_CHOWN Duration: 0.238 s
zeek-live-1          | find chown Duration: 0.316 s
zeek-live-1          | find chmod Duration: 0.217 s

Dashboards and netbox are the first to look at.

[mmguero]

I've made changes to dashboards and netbox and cut those down a lot:

just logs|grep Duration:
dashboards-1  | PUSER_CHOWN Duration: 0.005 s
dashboards-1  | find chown Duration: 0.982 s
dashboards-1  | find chmod Duration: 0.476 s
netbox-1            | PUSER_CHOWN Duration: 0.012 s
netbox-1             | find chown Duration: 0.783 s
netbox-1             | find chmod Duration: 0.395 s
zeek-1               | PUSER_CHOWN Duration: 0.249 s
zeek-1               | find chown Duration: 0.348 s
zeek-live-1         | PUSER_CHOWN Duration: 0.343 s
zeek-live-1          | find chown Duration: 0.405 s
zeek-live-1          | find chmod Duration: 0.213 s
zeek-1               | find chmod Duration: 0.195 s
logstash-1          | PUSER_CHOWN Duration: 0.005 s
logstash-1           | find chown Duration: 0.258 s
logstash-1           | find chmod Duration: 0.180 s

Going to look at zeek and see what's up there.