There are various combinations of running state for the daemon
process monitoring the collectd process and the process that it
is monitoring. In some pathological cases the original script
could hang for up to 60 seconds on 13.0 and indefinitely in 13.3

This commit addresses the problem by doing the following:

* consolidates the pre-stop and post-stop functions into a single
  stop function

* convert the rc script to use standard rc.subr functions to get pids

* validate that have pids we're trying to wait for

* Send a SIGKILL instead of SIGTERM to daemon process if the monitored
  process has gone away.

Detect critical failures earlier in test run and bail
to provide engineering with VM in failed state.

There's not really any point to proceeding with tests if our
attempt to create a pool fails.

Upstream FreeBSD changes to rc.subr break parsing of rc scripts
that have name variable containing a minus sign `-` in subtle ways.

wsdd fails to start in TrueNAS 13.3 due to configuration file
differences.

This fixes various situations in which the sharing.nfs.create will
fail when `hosts` is not specified.

This looks like a long-lived typo in the service config.

Several dependencies changed names and we no longer use py-zeroconf.

(cherry picked from commit 16c26350c078130ae72445505818774ee72ce5c6)
Co-authored-by: Caleb <yocalebo@gmail.com>

* Empty commit to create PR on github.

You should reset it

* change format of rsync push in module mode

---------
Co-authored-by: Stavros kois <s.kois@outlook.com>

Co-authored-by: M. Rehan <mrehanlm93@gmail.com>

NAS-126782 / 13.1 / PoolScrubService.__run: reduce memory usage during zpool history (by rkojedzinszky) (#12903)

We have deprecated this feature in Samba versions more recent than
4.15.

This parameter is deprecated and no longer exists in Samba 4.19.

Remove `test_exclude_recycle_bin` as it is not possible to run this test anymore without local S3 service (#12696)

Co-authored-by: Caleb <yocalebo@gmail.com>

The problem: The NFS bindip cannot be configured if the selected IP address is not functional. This setting is in /etc/rc.conf.freebsd. NFS is started and configured before the network is started. NFS has a post network start event handler that will restrart NFS if bindip is configured in the DB.

However, the NFS bindip setting was blocked at boot during the initial configuration run because the network subsystem had not yet completed startup. This resulted in no NFS bindip configuration.

The fix: In the post network start handler, regenerate rc.conf.freebsd before restarting NFS.

Co-authored-by: Caleb <yocalebo@gmail.com>

(cherry picked from commit 7a5ee4c7b3b57b861120c8117b7feb234560678b)

Decrease alert priority of zpool upgrade to NOTICE (#12299)                                                                                                                                                                                                                         Placing this as a warning encourages users to upgrade pool when                                                                           it's not really necessary and should be evaluated carefully by                                                                            the administrator. (#12300)
Co-authored-by: Andrew Walker <awalker@ixsystems.com>

Backend is required. This is set in SCALE, but change wasn't
backported to Core / Enterprise.

(cherry picked from commit e7399f10e1d042b27465d2e600e6ae0cd3f3ab66)
Co-authored-by: Caleb <yocalebo@gmail.com>

(cherry picked from commit 035e2d381c68053e5c0865147855d7f9c2fa713e)
Co-authored-by: themylogin <themylogin@gmail.com>

Co-authored-by: Caleb <yocalebo@gmail.com>

(cherry picked from commit 8b0e3e07b45ccc0f6f4a0ab41fe42110dc2122f5)
Co-authored-by: Caleb <yocalebo@gmail.com>