Note that this fixes the "current issue" but doesn't fix the underlying
problem. :/

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michele Tartara <mtartara@google.com>

scp by itself won't work on a cluster: a few extra options are needed to
tell it where to find the "right" ssh global known hosts file managed by
Ganeti.

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

This test uses a non-existing config value "file-storage-dir" and fails.
Fixing it by using the default value, which is currently what qa does.

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

The rename instance test performs an lvm-specific check for tags on the
instance volume. This should be in the future abstracted *inside* the
template class, and not explicitly called from qa.

This would allow other disks to perform the checks that make sense for
them. In the meantime though, we just make the same check only when the
backend storage for the instance is lvm-based.

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Michele Tartara <mtartara@google.com>

Setting up a proper environment for building a Ganeti can be quite difficult.

This patch adds a script for building a compressed chroot with all the
required libraries and tools.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

This patch rewrites and extends the design doc about storage reporting
with respect to disk templates and storage types. In constrast to the
previous version, we now consider disk templates as the user-facing
entity, that the user can dis/enable for the cluster. Storage types
on the other hand describe the underlying technology used by the various
disk templates. Storage reporting will use a mapping from disk templates
to storage types to pick the correct method to report the storage for
the respective disk templates.

Note that the design doc in this state still contains some questions and
FIXMEs. Feel free to comment on those. I will complete them directly or
in future patches.

Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

After changing the way gnt-xxx info print their output,
cli.FormatParameterDict() is not used anymore.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

The code is more modular, and the output is YAML-compliant. QA has been
updated to reflect the new format.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

The code is more modular, the output is more similar to "gnt-cluster info"
and is YAML-compliant.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

The new function can be used for group-level instance policies. Support for
roman numeral formatting has been dropped to make the code simpler. For
other info items it's already unsupported, and it's not exactly documented
anyway.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

This is based on the --donwgrade option of cfgupgrade, first introduced in
1709435e.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

The watcher should be disabled while upgrading.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

The version numbers of two programs had been swapped.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

This is required because of 0e79564a; otherwise, the pylint-qa target will
fail.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

Add a simple walk through the intended use case of
gnt-instance modify --new-primary for the disk template
file. The tested scenario is shutting down an instance,
manually moving the disk to the new node, setting the
new primary, and starting up the instance on the new
node. It is also verified that the operation is rejected
while the instance is still running on the old node.

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

Also perform QA tests for the file disk template. The main reason
is that testing gnt-instance modify --new-primary requires a
manual move of disks, which is most easily simulated for files.

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

Extend gnt-instance modify by a new parameter --new-primary
that sets the pnode paramter of the opcode.

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

If pnode is set LUInstanceSetParams now sets the primary node
accordingly. Unless force is set, it is verified that the instance
is no longer running on the old node.

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

This parameter will be used to set the new primary node
of an instance, assuming that the disks have been moved
by other means (outside ganeti's control).

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

Some required dependencies were missing from the installation instructions.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

Correct four typos in the comments of the Drbd types file.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Add status information as required by the design document.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Write an intermediate function computing the JSON data before adding them
to the DCReport.

Directly export the data computed by that function (with default parameters)
allowing them to be used (in the future) by the monitoring agent.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Also adds the DCStatusCode, part of DCStatus, and the addStatus
utility function for adding the "status" field to an already existing
JSValue.

The design document is updated to have the status codes sorted by increasing
seriousness.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Also, add it to the DRBD data collector, and export it from there.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Also, update the DRBD data collector to use and export it.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Also, update the JSON output (and the design document) so that it is not
in camelcase anymore. This is part of a bigger effort to remove camelcase
from the exposed JSON.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Define the new data type and update the DRBD data collector to use it.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Having 'null' instead of an empty string for an empty status message
increases the complexity of the code with no real improvement.

This patch modifies the monitoring agent design document solving this issue.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Additional clarification note added.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

A couple of them were missing and are added by this patch.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

The test is aimed mostly at instance policies, which changed recently.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

Instance policies in the configuration are upgraded and downgraded. Both
operations are idempotent.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

Minimum and maximum instance specs are put together into a single element
of the instance policy. This is in preparation for introducing multiple
min/max specs.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

It's now possible to go back to the previous stable version. Unit tests
provided.

This is mostly useful during development, when going from master to
stable/devel.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

All the upgrading code is now in one function. No functionality has been
changed.

Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>