With this commit, the DRBD data collector does not only print the data
it extracts, but includes them in the proper JSON structure common to
all the data collectors, as prescribed by the design document.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This commit adds the part of the JSON report generation code that will be
common to all the data collectors, according to the format specified in the
design document.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

The getCurrentTime function is of wide interest, so its moved to a more
appropriate position.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Currently, niceSort behaves differently than niceSortKey in cases
where the parsed groups are equal but the original not, as it uses
both members of the tuple as opposed to just the first one. Both
niceSortKey and the Python code use only the first one, so in corner
cases we can have different behaviour of these functions.

To rectify this, we change niceSort to be a simple wrapper over
"niceSortKey id", so that we get consistent behaviour.

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

This fixes a bug that corrups the config when one adds a
network to a node group. I wil soon extend QA by some commands
that would have detected this.

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

At the moment, because 'mergeData' is pure, it may set instance auto-repair
policies that are of the form `ArSuspended $ Until timestamp_in_the_past`.
If later on the auto-repair tool notices this, it has lost access to what
would be the next-in-line policy to use (and would have to re-parse all
tags again).

To avoid this, we pass the current time to 'mergeData' from ExtLoader.hs,
and we propagate it to Loader.getArPolicy. ExtLoader.loadExternalData is in
the IO monad, so it has ready access to getClockTime.

A few other places were calling 'mergeData' directly. For Hscan.hs and
IAlloc.hs, we add appropriate calls to getClockTime. For files under test/,
we use a current time of 0.

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

'getArPolicy' and 'setArPolicy' follow the precedence rules introduced in
b1eb71c7: within an object, the most restrictive tag wins; across object,
the nearest tag wins.

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Some auto-repair tags are composed of a prefix and then an optional
argument; the new "chompPrefix" function in Utils.hs allows to strip a
prefix, allowing the last character (the separator) to be absent if there
is no argument.

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This function allows to easily convert a ClockTime object to a string
representation of its timestamp (seconds-only).

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

AutoRepairPolicy, AutoRepairStatus, and other auxiliary types are added.
These are used only internally by the auto-repair tool, and parsed from the
various object tags as defined in the design doc.

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

In this commit, the AutoRepairType and AutoRepairResult types are defined,
with the possible values specified in doc/design-autorepair.rst.

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

A missing letter has been added to a docstring.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

This commits enables the DRBD data collector to use the Confd client to
gather information about the pairing between DRBD minors and instances.

For testing purposes, the DRBD data collector now requires either zero
or two parameters: one is the DRBD file, one is the pairings file.
When no parameter is passed, the collector is in "production mode" and takes
the data from /proc/drbd and from the Confd client.

The shell tests of mon-collector are updated accordingly.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This commit modifies the DRBD parser and its data structures to include
information about the instance a DRBD minor belongs to.

Test files have been updated as well, to support the new field in the data
structures.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Completion information for multi-command binaries were not generated in
the correct way, so the build-bash-complation script failed when such a
binary was actually found.

This commit fixes this issue, and a related docstring typo as well.

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This was extracted from a patch by Dimitris Aragiorgis with the subject
“Add locking to _NetworkQuery”. The rest of the patch did no longer
apply and will require more work.

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

If a received reply is valid and there is still no valid reply,
keep it, no matter how new it is.

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

While trying to understand why some code was not being tested, I
realised that we have a bad data type in Hcheck.hs.

We have "data Level = GroupLvl | ClusterLvl", but then we need to pass
the group name/index as well, so we have functions that look like the
following:

  f :: Level -> Maybe String -> …
  f GroupLvl (Just gname) = …
  f GroupLvl Nothing = … -- never used
  f ClusterLvl _ = … -- wrong, since a (Just x) could be passed

It's clear that we just need to move the group name in the "GroupLvl"
constructor, and simplify the call chain.

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

This removes one more tab conflict; this is the last module in our
code where we have both x.hs and x/.

Furthermore, we collapse all actual code into the new Main.hs module,
leaving the htools.hs basically empty (will allow better testing in
the future).

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

Per offline discussions, this is the first patch of the
renames. Tested with "make distcheck", seems to work fine.

The only change outside of the renaming is a bit of simplification in
the .gitignore rules; otherwise, simply s/htools/src/.

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

The modules are moved from the ‘top’ namespace to ‘Ganeti.HTools’, in
compliance with standard practices.

This patch moves the version string creation into a function in Utils
which shows some more information.

This patch adds a -V, --version command line option that shows the
program version and also updates the hn1 usage string (similar to hbal).

This just reorders some functions for a more logical ordering.

This patch changes the solution list to include a step counter so that
it's more clear these are successive steps (in a definite order), and
not just an unordered list of changes.

This patch changes the gnt-instance failover to migrate, and fixes a bug
in the formatting of commands.

This patch added a verbose output and changed the output so that by
default it is less verbose and more clear.

This patch adds a new instance move, FailoverAndReplace, which promotes
the old secondary to primary and then uses a new secondary node.

This is the last move that we can do within the limitations of one node
changed per move.

This patch moves the formatting of the command list to Cluster.hs and
enhances it with separator messages between the steps.

This patch adds a new replace secondary and failover move (equals to
“r:x f”), which can improve the solution (since we are testing more
options at each step).

Aligned the comments in Instance.hs

Currently hbal does multiple rounds, stopping when a rounds doesn't
bring improvements. With the recent changes to not remove instances from
the candidate list, this is obsolete as the first round will always run
to the end of the improvements.

This patch changes this so that the Cluster.checkMove function doesn't
recurse, but just computes the next best move (as its docstring says).
This means we can actually incrementally compute and print the solution,
and this is needed as otherwise an instance could move twice and the
second time it needs the current placement to compute the exact command
line and operation needed for the move.

This abstracts the individual placement solution so that it can be used
independently.

The current code restricts each instance to one move per round. This is
bad, as an computation restarted in the middle of the solution will have
a different set of instances to work and will thus lead to a different
end-solution.

Once this is applied, further rounds are not possible since the first
round will have tried all instances at its end. As such, the removal of
the rounds feature will be next.

The code adds a hard-coded 100 moves limit, which for big clusters is
actually small.

This prints the total memory/disk and also adds a header.

This is to keep in style with the other memory variables.

This is useful and not easy to compute otherwise.