This enables the per-node load/total available capacity scores to be
used in balancing. Note that the total available capacity is currently
fixed at zero and cannot be changed by the user.

This patch adds loading and processing the utilisation data during
instance moves. While the data is not yet used, it is correctly modified
by instance changes between nodes.

hbal has the new ‘-U’ command line argument for this. The format of the
file is simply instance name and the four stats, space-separated.

The latter is only used right after the former in the Loader module, and
we'll need more of this 'update not with the data of this instance'
functionality (which is different than addPri where all information must
be updated).

The patch also changes the signature of Node.setSec (to remain
consistent with setPri).

These were already duplicate (Text and Simu) and we need tryRead in more places.

The strange printf usage is due to some limitation (it seems) in ghc for
very long argument lists. The whole printout should be rewritten later.

This patch adds the datatypes and modifies the nodes and instance types to have
such attributes. They are not used yet in any way.

This is similar to --print-nodes, but with much fewer fields.

This exports all functions, but it's still good to have.

This makes (for now) the code hlint-clean. This is per se not a huge
gain, but it allows easier tracking of regressions in style later
(one-two new violations are easier to diagnose when not hidden among 20
“known” ones).

This is again the cs_x to csX name change.

This changes from a_b to aB in all node and instance attributes, to
match the standard Haskell style. Also attributes that should have been
camel-cased but weren't were changed (e.g. plist → pList, pnode →
pNode).

Before we used a tuple; since we'll need more metrics in the future,
it's simpler to transform this into a list of doubles, whose elements
are handled homogeneously by all the code that needs them.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

This not so nice patch adds submit/wait-for-completion functionality in
hbal. When enabled via ‘-X’ (only working on Luxi), it executes the jobs
(as split into separate jobsets by Cluster.splitJobs) and waits until
ganeti has finished all of them.

This is a rough version: it waits ad infinitum for completion, has
hardcoded timeouts, etc.

Signed-off-by: Iustin Pop <iustin@google.com>

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

This are higher level wrappers over the basic callMethod.

Signed-off-by: Iustin Pop <iustin@google.com>

The current Cluster.iMoveToJob always creates failovers, which is not
what we want. This simply used the original instances status to select
between these two (this is not optimal by the way, since the status
could have changed in the meantime).

Signed-off-by: Iustin Pop <iustin@google.com>

This will be needed in order to generate the proper instance move commands.

Signed-off-by: Iustin Pop <iustin@google.com>

This holds for now just the job status definitions and its serialisation
to/from JSON.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Since on regular compiles the version information is not critical, it's
better to only update it on make dist; furthermore, there it is indeed
needed to regenerate it automatically, without requiring the use of make
maintainerclean beforehand.

This patch simply adds a new phony target that forces regeneration as a
prerequisite for the dist rule.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

It seems that haddock cannot document tuple members - but arguably, once
one needs to do that, tuples should not be used anymore.

This just moves the comments to the tuple comment.

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

If we're building on a different version, the version file needs to be
regenerated.

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Currently parseNode skips looking for values in offline nodes, but tries
to read them for drained ones. With this patch we treat offline and
drained nodes in the same way (which is compatible with the iallocator
protocol changes introduced in Ganeti 2.0.4, but also doesn't break
retrocompatibility). As a bonus we don't need to calculate the value of
the internal Node "offline" parameter, because we know that the node is
going to be online anyway, if we lookup values (and it was statically
set to offline, before, in the offline case).

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

This will automatically sort our Ganeti jobs into the independent job
sets, and then we can submit them separately.

This converts from htools-specific Placements into Ganeti standard
OpCodes, which will later allow execution via Luxi.

These are just a few opcodes we need for executing instance moves.

This will allow a more descriptive output later in the solution list, as
opposed to trying to reconstruct the move from the node indices.

The patch also documents the Placement members.

The Luxi loader implements both a generic Ganeti Luxi client and the
loader; it is better if these two are separated. The patch adds a
Ganeti/Luxi.hs (not under HTools!) since that is generic for Ganeti, and
not related necessarily to htools.

Since moving two instances between different node-quadruples (inst X: A,
B → C, D and inst Y: E, F → G, H) can be parallelised by Ganeti, it
makes sense to split the operation list into jobsets whose execution
must be serialised, but whose individual jobs can be parallelised.

Of course, this doesn't help, because today a single logical job is for
some cases split in multiple gnt-instance invocations, which (if run
using ‘--submit’) might be wrongly parallelised by Ganeti. So while for
now this is mostly a cosmetic hint to the user, and it just a step for
future improvements.

Due to the Control.Exception changes between 6.8 and 6.10, using it
portably is difficult. Since we're only interested in handling I/O
errors, we can use prelude's catch and not have to deal with
Control.Exception at all.

The use in Luxi.hs where we just use bracket is fine across the two
versions, so we keep that.

Haddock doesn't like pre-processed files (at least not in all versions).
Thus we need to remove the ExtLoader module from the haddock-procesed
file list.

Test.QuickCheck.Batch.runTests doesn't return any error statistics,
which makes the test suite just display errorrs and always exit with
exit code 0. This is not good, since one cannot then actually batch run
tests.

This patch adds a wrapper over Batch.run which also modifies a passed
IORef Int to keep track of how many test failures or test aborts we had.
This makes it easy then to exit with an appropriate exit code.

The Makefile was intented to be -Wall and not simply -W, but I missed
that. This enables more warnings and also enables -Werror (except for
the tests).

During testing I used the test inversely to see it triggers correctly,
and committed by mistake the inverted test. Fixing it.