Aggregate all relevant cluster info into single endpoint for use by the TrueCommand team.
This PR also alters returns for relevant `ctdb.status` to be results of call to new endpoint
(cluster.management.summary). This should simplify work required for clustering consumers.

There is minor descripancy between ValidationError and
ValidationErrors regarding default errno that is set. It makes sense to
use EINVAL in both cases since they're related to validation failures.

This commit adds basic support for clustered SMB accounts.
Address space is carved out of the high end of what is
supportable for idmaps in general. Validation for AD idmap
settings is tweaked to prevent users from carving into this
range.

The accounts have are disabled via shadow file, but have a
valid NT hash for SMB access. This allows them to be used
in situations where clustering user needs non-AD clustered
SMB access, a primary example of where this may be useful
is for micro-benchmarking cluster performance in an
automated manner. The accounts are not valid for SSH, shell,
or API access.

CI tests are also added to validate that clustered users
can access SMB shares via any node's public IPs and that
they cannot access shares if they have been administratively
disabled (locked).

* Use netdata for retrieving disks temperatures

* Remove nvme util from middleware

* Properly fix nvme class

* Do not read nvme temp twice

This commit adds a separate endpoint to validate pool name so UI can consume it at the start of pool creation wizard and error out before user proceeds to next step in the wizard.

This commit adds changes to ignore any interfaces created by tailscale app when running in host network mode as they are not managed by middleware and will result in all sorts of different issues.

Attempts to access ZFS ctldir via libgfapi can cause gluster
volume mount to freeze. We should prevent this variant of
foot-shooting by adding robust checks to ensure that only
directories within ZFS datasets are used. Since we already
have filesystem.stat output as part of brick validation we
can easily expand existence check to also cover things like
fstype, whether the brick path is empty, dataset mount options,
etc.

This test intermittently fails with an error that another
transaction is in progress. Since it can cause a cascade of
other test failures, we should just retry a few times before
failing. The error is probably spurious due to VMs that we
are running the cluster on.

As we deprecated the ctdb shared volume, the migrate method was
moved from ctdb.shared.volume namespace to ctdb.root_dir namespace.

Add some basic validation of payloads delivered related to brick
and volume type, and expand cluster tests to validate that
volumes are created properly.

Testing for validation and normalization of gluster config, and
creating different volume configurations is achieved by using
multiple bricks on the same server. This is a rough analogue to
supported cluster configurations and something we can address
later as the testing infrastructure for automation improves.

* Fix proftpd startup after reboot

There are two options for running proftpd: standalone and under xinetd.

The default method is standalone and is the method we use.

The xinetd option runs 'proftpd.socket'.
The default standalone runs 'proftpd.service'.
Both cannot be running at the same time.

Somehow, the debian packagers allowed proftpd.socket to start
in the default standalone configuration.  This condition blocks the
start of proftpd.service and is the source of this error condition.

* Add CI test for NAS-123024: FTP service starts after reboot. (currently disabled)

This commit adds changes to not spam middleware logs again and again if the error has already been logged once.

There are some situations in which the clustered services
file may exist, but is empty. This should be treated in the
same way in the ctdb event script as the case where the file
simply doesn't exist (load our defaults).

* Prevent logged in iSCSI targets being included in get_disks

* Add is_iscsi_device test to retaste_disks_impl

* Add test_24_iscsi_target_disk_login

For a regular system, this will perform a loopback iSCSI login using
the public IP.  For a HA system it will create an ALUA iSCSI target
and check the standby node (which is logged into the hidden internal
iSCSI target).