GitLab

This patch adds checking of cluster data in the binaries and display of
node's x_mem/i_mem in the node list.

This patch adds a checkData function which goes over the node list and computes
the unaccounted memory, returning a list of warning messages (if any) and the
update nodes.

This patch adds a new n_mem field to the node objects, and implements
read/save/show support for it. The field is not currently used (except
in the node list) but will be used for checking data consistency and
instance up/down status.

This patch changes the parameters passed to the node and instance
constructors from generic Strings (which are then parsed via “read”) to
the actual used types, by converting them earlier in Cluster.loadData.

This patch does some small changes:
  - fixes a comment
  - export more node functions (unneeded now, but hscan will use them)
  - fixes Makefile rule for building the programs

This is added for better readability, since this is very often used in
declarations.

This patch changes two things with regard to offline nodes:
  - first, it only calculates the various coefficients across online
	nodes
  - second, it adds a new score denoting the percentage of instances
	which live on such nodes

The first change allows correct score computation in presence of offline
nodes (whose properties we don't need to take into account), while the
second change actively evacuates offline nodes.

This patch adds a new ‘-’ flag for the node status which denotes offline
nodes.

This patch changes the Cluster.checkInstanceMove function to restrict
the target move list based on which nodes are online.

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

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 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).

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.

Since for a very many N+1 failures in a cluster, we could actually
degrade the N1 CV by making a node N+1 compliant, we need to make sure
this value only decreases when fixing non-compliant nodes.

The easiest way is to compute the N+1 score as a percentage of failed
nodes, with the caveat that the domain of values might not be fully
compatible with the other scores. It is still [0, 1] but does not vary
like the others.

This patch adds two new variables to the cluster score:
  - variance of the failN1 attribute
  - variance of the reserved memory percentage

The variance of the failN1 helps make the cluster N+1 happy, whereas the
reserved memory percentage helps balance the unused memory for
redundancy on the nodes.

This patch adds a score variable to the placement type, so we can record
the changes in the cluster CV for later display.

This gives visibility in the decrease of the parameters and can show
which are the most important steps to perform (out of the full move
list).

This patch removes the primary/secondary instance lists from the node
status and also removes the tabbed formatting with explicit width
formatting.

This patch makes the tabular solution list nicer, by changing from tabs
to explicit widths.

This patch automatically removes the longest common (domain, i.e.
starting with a dot) suffix from the node and instance names. This gives
a much clearer display, and this format is compatible with the way
Ganeti accepts shortened names.

This just makes indendation nicer in many expressions.

This patch changes the computation of p_mem / p_dsk from on-demand
(whenever the cluster stats are computed) to after-modify (after a node
is modified, we update its stats). This brings a god speed-up as only
one node or two are usually changed between cluster-wide stats are
computed.

This patch flattens the two folds into one, by simply building the whole
list of moves instead  of the double recursion (nodes and the each
node's moves). This has no functional change, but it's much cleaner.

Currently we always compute the available node list for moves (for an
instances) based on the nodes of the initial table. This works find,
however is a repeated calculation.

We optimize this by passing a node list (of indexes, not full objects),
which helps in two ways:
  - faster to filter later
  - allows restriction of target nodes by enforcing only this subset as
    target for moves

This cleans up and splits the individual instance move into a separate function.

This patch changes the balancing algorithm to not iterate linearly over
the instances (in a random, but fixed order), instead selecting at each
step the best next move. This should allow a better score (most of the
time), and usually also a shorter solution.

We need to lookup the moving accumulator, of course.

This patch changes the format of the input node and instance lists. It
removes the list of primary and secondary instances from the nodes, and
adds the primary and secondary nodes to the instance list.

This is done so that we can have the same input data from commands as
from the RAPI.

This is the initial import of release 0.0.3.