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

This enables unittesting of RAPI resources.

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

This base class, which employs a meta class for the actual work, allows
easier definitions of RAPI resources using opcodes. Follow-up patches
will change some of the existing RAPI resources.

The long-term goal with these changes is to make it easier to verify the
consistency of RAPI with CLI (or the underlying opcodes).

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

Apart from making match Ganeti's code style the name is also more
descriptive.

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

Unlike stand-alone functions these will be relatively easy to override
for unittests.

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

They were only used in one place and upcoming changes, enabling testing
of resource classes, will be easier with this merge.

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

These were never really useful. Neither gave a complete list of
available resources—the documentation in doc/rapi.rst is much better at
that.

Since some monitoring code might use the “/” resource it's kept around
but will only return a JSON “null”.

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

Signed-off-by: Tsachy Shacham <tsachy@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Agata Murawska <agatamurawska@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Agata Murawska <agatamurawska@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Turns out my attempts to get it work nicely with “make distcheck” broke
the detection and of course I only notice after pushing.

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

This check will abort if one of the standalone modules (currently
lib/rapi/client.py and tools/ganeti-listrunner) imports anything from
the Ganeti source directory.

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

* devel-2.5:
  Delete master IPs from mergee master nodes
  Use pep8 utility in “make lint”
  Two more PEP8 fixes
  check-python-code: Give location(s) of lines longer than 80 chars

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

Added a step in cluster-merge that removes the cluster IP from the
master node of the mergee clusters.

Signed-off-by: Andrea Spadaccini <spadaccio@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

This utility checks whether the code conforms to PEP8. Some checks had
to be disabled for Ganeti.

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

cmdlib: Avoid wrapping using backslash

gnt_group: Avoid ** magic using keyword arguments (the “pep8” tool
doesn't like the inline comment in this case and will complain about
spaces around the “**” operator)

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

Until now it would only say that there was a line longer than 80
characters, but not where.

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

Signed-off-by: Tsachy Shacham <tsachy@google.com>
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

* devel-2.5: (32 commits)
  PEP8 style fixes
  Wrap a few long lines
  listrunner: Avoid exception if machine is rebooted
  Remove wrong type declaration from option
  Fix wrong method name in cluster-merge
  Fix --skip-stop-instances help message
  cluster-merge: Add the --skip-stop-instances opt
  watcher: Wait for child processes by default
  Update release date in NEWS for 2.5.0~beta2
  Try 3 times before giving up on per-node commands
  Allow retrying commands in cluster-merge
  Add a TODO on the VerifyCluster option
  Transform node readd exceptions into just errors
  Offline node when adding it to a merged cluster
  Bump version to 2.5.0~beta2
  sphinx_ext: workaround epydoc warning
  check-news: Show per-file line number
  Unify some file headers
  Makefile: Add design-ovf-support to list of doc files
  ensure-dirs: Fix epydoc error
  ...

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

Identified using the “pep8” utility.

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

Signed-off-by: Tsachy Shacham <tsachy@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Tsachy Shacham <tsachy@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Tsachy Shacham <tsachy@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Tsachy Shacham <tsachy@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Had to break it as well, today! ;)

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

Signed-off-by: Agata Murawska <agatamurawska@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

… provided that disk_template value is set in the config.ini file.

Signed-off-by: Agata Murawska <agatamurawska@google.com>
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

This saves another 5 kB from the generated script. An old debug comment
written to the generated output is also removed.

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

Handle exceptions gracefully when trying to read the command's output.

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: Andrea Spadaccini <spadaccio@google.com>

Fixed a wrong method name in the last patch.

Signed-off-by: Andrea Spadaccini <spadaccio@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

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

This option allows to do a check for running instances on the mergee
clusters instead of stopping them.

Signed-off-by: Andrea Spadaccini <spadaccio@google.com>
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

This patch retains the behaviour of ganeti-watcher in previous Ganeti
versions.

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>

When contacting lots of nodes some may fail. Give it a couple more
chances before giving up on them.

Possible future TODO: continue, but just mark them as offline.

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

Signed-off-by: Andrea Spadaccini <spadaccio@google.com>
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>

We are after the point of no return, no point in failing everything
because a node failed to readd. Better to just report it and move on.

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>