Make hroller output the node groups not containing the master node
sorted by size, largest group first. The master node still remains
the last node of the last reboot group. In this way, most progress
is made when switching back to normal cluster operations after the
first reboot group.

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Thomas Thrainer <thomasth@google.com>

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

Add option --node-tags to tell hroller to consider only nodes
with these tags. A use case would be a tag tracking on which
nodes the maintenance has not yet been carried out, e.g., if
rolling reboots are interleaved with other cluster operations.

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

Since the htools representation of a node now allows adding
the node tags, populate this field correctly in the Rapi
backend.

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

Since the htools representation of a node now allows adding
the node tags, populate this field correctly in the LUXI
backend.

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

In order to allow htools to make use of node tags, add them to the
text format. This is done by adding a new column at the end of the
node lines. If this column is missing, the default value (which
is the empty list) is left unchanged, thus yielding the current
behavior.

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

Since hroller (and probably other tools in the future) will support
node selection based on node tags, extend the node data structure to
allow adding this information.

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

Hroller used to first compute a coloring of the node graph and then
filter out the nodes that it had to work on. While the only filtering
was according to node groups this did not make a difference, as there
shouldn't be any instance with primary and secondary node on different
node groups. With more elaborate filtering, however, reducing the graph
first can lead to better reboot groups.

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

Change the behavior of mkNodeGraph to tacitly ignore all instances
where one of the nodes is not in the list of nodes. In this way, we
can construct sub-graphs by filtering the nodes and ignoring any
possibly added isolated nodes for the missing indexes.

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

The haskell type definition of opcodes should remain aligned with the python
one.

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

It will be added to the haskell definition of opcodes, to keep it aligned to
the python one, and it will be used for fetching the reason trail by the
instance status data collector.

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

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

Produce a personalized 404 error when the requested resource is not available.

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

Implement the API function of the monitoring daemon that provides the report
of all the data collectors.

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

Allow to ask the monitoring daemon for the report of one specific data
collector.

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

Export the full report instead of just the data from the DRBD data collector.

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

Change the JSON serialization for the "category" field of data collectors,
in accordance to the latest version of the design document.

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

Implement the handler for outputting the list of collectors
(name, category, kind) in JSON format.

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

Since we use the primitive string type for group UUIDs, the group
fields have a bug where we pass the group name as filter for node
tests, whereas the nodes themselves use the group UUID. This results
in zero node count, empty node list, and no instances being reported
as assigned to groups.

The patch fixes this and adds a test for the node count. It does some
test generation improvement, and also cleans up whitespace issues in
Test/G/Q/Query.hs (the functions case_queryNode_allfields,
prop_queryGroup_noUnknown and case_queryGroup_allfields are unchanged
but simply have indentation fixed).

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

Cherry-pick of e7124835, fixes issue 436
Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

Conflicts:
	test/hs/Test/Ganeti/Objects.hs
	test/hs/Test/Ganeti/Query/Query.hs

With tiered allocation, hspace uses all the max specs in turn as the
initial instance spec.

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

Now instance policies can contain more than one min/max specs.  This is the
main element of the "Constrained instance sizes" section in the
"Partitioned Ganeti" design doc.

This is a big patch, but changing the type of a configuration item requires
to change all the code that handles it.

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

So far, any successful query was only logged at debug level, thus leaving
no traces in the logs, unless debugging was enabled. Therefore, we add
a notes about successful completion at info level. Also, repeat a failed
request at WARN level, so that, even if debugging disabled, the request
is available together with the error message in the logs.

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

Specs used to specify limits in instance policies are always complete, even
at group level (only the whole limit set can be overridden). This is in
preparation for introducing multiple limits.

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

In a simulated cluster as created by the simulation backend to
the htools, make the first node of the first node group the master
node. In this way, hools (like hroller) that require a master node
can also be used on simulated clusters, e.g., for testing.

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Michele Tartara <mtartara@google.com>

Sometimes, e.g., for hroller, it is necessary to know which node
is the master node. Therefore this information has to be included
in the text format as well. Since we never use an offline node as
master node, we can put this information can be put in the "is
offline" field; as old versions only check this field for equality
to "Y", using "M" to indicate the master is compatible in both
directions. This also leaves room to add more roles in the future,
should we need them.

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

Remove the useless parts of the old, partial, implementation of the support for
tracking the reason of instances state change, before implementing the new
reason trail support, as per the design document.

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

As people rely on the master node being the last node of the last
group, make hroller fail, if no master node could be found in the
cluster. This happens, e.g., if a backend format is used that does not
contain information about the master node. The error can be turned
into a warning by using the --force option.

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

The cluster data contains the information about the master
node. Use this information to set the isMaster bit correctly.

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

The cluster data contains the information about the master
node. Use this information to set the isMaster bit correctly.

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

The information about which node is the master node is a cluster-wide
setting, in most formats provided independently of the node information.
Most backends therefore have to set the isMaster bit indepently in the
obtained node list. This patch provides a utility function for this.

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

If in the list of nodes to be scheduled for maintaince,
one is marked as being the master node, schedule it
as the last node in the last group.

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

Extend the description of the node by the property
of being the master node; also provide and appropriate
setter function. This property will be used, e.g., by
hroller to schedule reboot of the master last.

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

These lines are ok according to previous versions of hlint but trigger
an error with version 1.8.43.

Signed-off-by: Thomas Thrainer <thomasth@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

This commits adds name to NIC and Disk modification definition.

Signed-off-by: Chris Stavrakakis <cstavr@grnet.gr>
Reviewed-by: Helga Velroyen <helgav@google.com>

Add name and uuid fields to Disk and nic Haskell objects. Also,
since they now have an UUID, make them instances of UuidObject.

Signed-off-by: Christor Stavrakakis <cstavr@grnet.gr>
Reviewed-by: Helga Velroyen <helgav@google.com>

Signed-off-by: Klaus Aehlig <aehlig@google.com>
Reviewed-by: Michele Tartara <mtartara@google.com>

Since storage type are supposed to represent the underlying technology
of disk templates, the storage type 'sharedfile' is superfluous, because
technically both disk templates 'file' and 'sharedfile' use the file system.
This will be of use when implementing the storage space reporting for
all disk templates.

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

Since managing of different storage units is now done using
disk templates and not storage types, we remove the obsolete
enabled storage types.

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

Adds setting of the list of enabled disk templates to 'gnt-cluster modify'.
Note that this does not yet include any verification regarding disk templates
currently used by instances.

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