Use a function instead of a variable written by another module.

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

This avoids having to monkey-patch the utils module for unittests.
Monkey patching is evil and caused a bug while moving code around.

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

Sorry I thought I did run commit-check but must not have paid attention
to its output. There was a typo in the docstring. This patch fixes this.

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

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

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

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

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

The hooks is due to me not checking the recent patch from David, and
the cli.py change is due to the utils.py split.

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

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Emit a warning for each instance with a mirrored disk template that has
its primary and secondary nodes living in different node groups.

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

This operation allows to assign nodes to a group, verifying whether
the change would leave any mirrored instance split between different
node groups.

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@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>

This makes it a public function, which is necessary for
moving code into separate files.

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>

The unittest file is renamed to match the other utils.* tests.

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>

This is preparing for splitting the whole units module.

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

While reviewing dato's interdiff for the OpAssignGroupNodes, I
realised that we can do better. This patch replaces the hand-built
DISPATCH_TABLE with one built from the opcode.OP_MAPPING dict.

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

Currently, the unittest TestDispatchTable in mcpu unittest does a
hard-coded approach to test whether an opcode should be included or
not in the mcpu.Processor dispatch table. This is not flexible, so we
replace it with two changes:

- first, we do not return the base OpCode in opcodes.OP_MAPPING;
  rationale being that we shouldn't need to serialise or deserialise
  this opcode during normal operation
- second, we add a new attribute WITH_LU (other name suggestions
  welcome) to the base opcode set to True, and we explicitly set it to
  False in OpTestDummy, thus automatic skipping of such LUs in the
  unittest (and in other places)

To correct for the new behaviour, the tests in the opcode unittests
are updated to include OpCode explicitly.

Note: we also replace assert_ with assertTrue in the affected unittest
(as assert_ is deprecated in 2.7).

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

Currently, the IAllocator doesn't know that non-vm_capable nodes
should not be queried for hypervisor/instance information.

This patch updates the node data computation to generate basic info,
but not the runtime data for such nodes. Furthermore, the
documentation is updated to note that such nodes will node have
runtime data.

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

This makes use of the new way of returning console information from the
master daemon. Unittests are included.

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

With this patch OpConnectConsole will no longer just return a command
with arguments, but rather a detailed description about how to connect
to an instance's console. Unittests for some parts are included. Another
patch will follow to adjust the hypervisor abstractions for the new
model.

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

If the SSH command fails, this will give a more detailed error
message than before.

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

The source cluster has to shut down an instance before it can be
exported. Doing so can take a while, but the default connection timeout
is only 60 seconds. Adding the shutdown timeout on the receiving cluster
should help.

Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
(cherry picked from commit dae91d02)

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

This field is based on OOB support and is only available if there's oob
support for that node.

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

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

_GetQueryImplementation() uses _QUERY_IMPL, which list all query type
implementations. By moving it to the end of the file, we ensure all such
classes are defined by the time we list them.

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Additionally, change TestQueryFields.testSomeFields() to handle lists of
fields shorter than 20 elements.

Signed-off-by: Adeodato Simo <dato@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Recently, we've seen more and more cases of a specific breakage
pattern in Ganeti: master candidates which are semi-alive (as in, they
respond to ping, they can complete a TCP/SSL handshake, but otherwise
the root filesystem is broken) cause lots of confusion within masterd.

My analysis shows that waiting up to 5 minutes for a reply from such a
broken master candidate is too long, and this long wait breaks other
timeouts (e.g. the Luxi timeout), making standard recovery from this
situation very hard. It's much easier to kill the master daemon, edit
manually the config file and mark the node as regular, then restart
the master daemon.

The proposal is therefore to reduce the timeout for the job queue
functions to TMO_URGENT (1 minute), which should be more balanced
between a working but overloaded node and a broken node.

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

Locks can now be queried using “Query(what="lock", …)” over LUXI.

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