Together with Michael we refactored the code to make it better and
easier error reporting. Without printing backtraces for authentication
and verification issues.

Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Some directories don't exist in the repository, but are required at build time
(e.g. doc/html). Until now some were created explicitly, some through the
target “stamp-directories” and other target simply relied on a previous target
to create the directory.

This patch tries to clean this up by getting rid of “stamp-directories” and
instead use rules to recreate any missing directory. As described in a comment
in the code, a file inside each directory is necessary, named “.dir”.

Order-only dependencies are used for directory creation to avoid rebuilding
where only the “.dir” file is missing (see “info make”, section “4.3 Types of
Prerequisites”).

The target for building the documentation is also changed to use “…/index.html”
instead of a hidden file. Some style changes are also made.

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

A new constant, LUXI_VERSION, is used to verify the peer's version. The
version is optional, so old(er) clients and servers talking to peers not
supporting it won't break. Example with mismatching library:

$ gnt-instance list
Unhandled Ganeti error: LUXI version mismatch, server 2020000, request
1010000

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

This allows LUXI errors to be encoded and serialized.

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

To remove the instance after an export it needs to be stopped. This can
be achived using the parameter “shutdown”, or by explicitly shutting
down the instance before exporting. The latter would still require the
“shutdown” parameter to be set. To make it more intuitive, this
requirement is changed with this patch. Instances already stopped are
accepted for automatic removal.

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

On my machine it takes over 30 seconds, disabling it can
speed up the QA.

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

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

The nodes and instances parameters to the constructor are mandatory
anyway, as a value of None will fail when creating the LockSet. Rather
than fixing this adding code lines, since we never used the default
value, let's remove them and require that the parameters are passed.

This also fixes the only places where we inited GanetiLockManager with
keyed parameters and without arguments.

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

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

In particular:

  - introduce a "gnt-group" command to hold group-level operations.
  - ditch the concept of "default node group", except for single-group
    clusters.
  - introduce an "alloc_policy" attribute for node groups, indicating
    how they should be treated by automated allocation tools.
  - introduce a "drain" operation on node groups.
  - define iallocator modes for new instance allocation and
    inter-group moves (choosing among all groups, or providing a
    limiting list).
  - indicate and explain that changing the group of a node will be
    initially only supported for nodes that are empty.

Signed-off-by: Adeodato Simo <dato@google.com>
Signed-off-by: Balazs Lecz <leczb@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

This is just a basic check, plus a warning. In the future, we might do
more checks, or prevent simple onlining (without readd) if --force is
not passed.

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

We will need the new role in CheckPrereq, so move its computation there
and save the new role to self.

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

This small patch modifies LUCreateInstance, LUReplaceDisks and
LUMoveInstance to not use non-vm_capable nodes.

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

Also changes the error code for the other CheckNode* helpers to
ECODE_STATE, not ECODE_INVAL: ECODE_INVAL is for requests that are
invalid (e.g. create drbd instance with one node), whereas ECODE_STATE
denote requests that are not satisfiable due to cluster/node/instance
state.

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

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

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

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

This is used in two places already, and will be needed in a third, so
let's abstract it.

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

The method to make vm_capable integrate easily into cluster verify is as follows:

- we add a new NV_VMNODES that represents *non*-vm-capable nodes
- the LU populates this list (it's expected that non-vm_capable nodes
  are few compared to vm_capable nodes)
- backend skips the checks that are related to VM hosting
- in the LU, we reorder the VM-related checks so that they occur after
  the non-VM (generic) tests, and we only execute them conditionally

Additionally, we add some support to the instance checks to detect
instances living on bad nodes.

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

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

And also do some cleanup: we only run the role changed actions if the
node has actually changed roles.

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

This can be used to compute a node's instances easily, and a small
function to get all non-vm_capable nodes.

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

This will show a warning if, for example, one side of a DRBD
disk becomes unavailable. The data is collected separately
from the other verification data.

Example output:

* Verifying instance status
 - ERROR: instance inst1: disk/0 on node2 is faulty

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

- run-in-tempdir should depend on what it copies to the temporary dir
- Add PYTHON_BOOTSTRAP to BUILT_SOURCES
- Don't use “mkdir -p” directly
- Create directory if necessary for writing bootstrap script

In summary, this should make “make distcheck” in pristine
checkout work again.

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

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

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

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

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

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

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

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

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

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

Eventually this will help ensuring that clients and servers are of the
same version, as long as they're imported from the same path. Currently
it's relatively easy for gnt-* and ganeti-* to be from a different
version.

Scripts will be at ganeti.client.gnt_* and a small bootstrap script
calls a “Main” function from the module.

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

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

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

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

This adds the flag and some initial handling. The rest of the changes,
for cmdlib, come in a separate patch.

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

There have been many bugs in gnt-node modify. Let's try to introduce
some more.

This patch reworks the node role changes from tracking the flag changes
to completely overwriting the flags based on the new role. This paves
the way for (in 2.4 or later) moving to a single attribute for nodes.

We compute the old role and the new role based on the required changes
and whether we need to auto-promote. Once this is done, the body of the
Exec() function becomes trivial (there's more code related to output
formatting than the node flag changes).

Another advantage of the new version is that the entire flags are
overwritten, and that all are changed at the same time, making it
impossible (harder?) to have partial updates.

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

Signed-off-by: Balazs Lecz <leczb@google.com>
[dato@google.com: extracted language fixes from bigger patch.]
Signed-off-by: Adeodato Simo <dato@google.com>

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