Since we use the cluster name for the SSH known_hosts file, ssh will
always detect a changed IP (since we never connect to the cluster master
name, but the node names), and will complain about it/try to update the
user known hosts file (since that is /dev/null, it doesn't matter, but
it's not nice). So we disable the IP check.

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

utils.py, where they were before, is already huge.

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

This variable can be used by other tools to determine in a generic way
whether a given hypervisor supports migration or not.

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

… since the opcode doesn't auto-initialize to None.

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

This allows cleanup to be done properly if the "instance" wasn't running
at all (based on the CleanupInstance framework, instead of the retry in
hypervisor).

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

Commit 9cf4321f indented some lines by
4 spaces rather than 2, and was git-amed without noticing. Fixing.

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

The function's code was mostly copied from Python's documentation
and it's equivalent to “functools.partial” in Python 2.5 and above.

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

boot=on implies that KVM boots using extboot. This is only required
to boot non-IDE disks and has the side-effect that there is at most
one bootable device. This behaviour breaks some operating systems,
most notably the windows installer that tries to chainload the hard-disk
from the CD-ROM after the initial part of the installation.

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

When using VNC, it is recommended to use a tablet-style
input device instead of a mouse. This allows most VNC viewers
to send proper mouse coordinates to the virtual machine's desktop
resulting in perfectly aligned guest and host mouse pointers.

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

If the uid pool is not in use _TryReadUidFile will try to return "uid"
even if it hasn't tried to read it at all.

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

Ensure that all RAPI client methods that should return job ids actually
return job ids.

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

In order not to complicate to much the _ExecuteKVMRuntime function with
nested try/except/finally/else constructs we move the actual
runcmd+check call in a separate _RunKVMCmd function.

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

If this file exists, for an instance, we release the uid contained
inside back to the uid pool, at instance shutdown.

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

We can't guarantee to keep the LockedObject around, when ReleaseUid is
called (because that might happen, for example, after a ganeti
shutdown/restart). So we'll support releasing on just an integer uid.

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

uidpool users that get passed a LockedUid by the uidpool need to know
which one the uid is. Since it's not nice to make them access a private
member, and we shouldn't make "uid" public either, we'll add a getter
method.

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

Currently putting the cleanup just after the actual shutdown ensures
that it never gets called, due to how the retry/shutdown cycle works.
So we move those cleanups to their new dedicated place.

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

Currently some hypervisors (namely kvm) need to do some cleanup after
making sure an instance is stopped. With the moving of the retry cycle
in backend those cleanups were never done. In order to solve this we add
a new optional hypervisor function, CleanupInstance, which gets called
at the end of the shutdown procedure, and which interested hypervisors
can implement to be sure not to miss cleanup operations.

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

When enabling a new hypervisor, we must ensure that we have its
parameters in the cluster (global) hvparams dictionary.

Furthermore, we add a verify check for this case (this should be better
replaced with an auto-fix?).

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

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

Signed-off-by: Balazs Lecz <leczb@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

The primary IP is not used for DRBD disks, only for the job queue and
internal commands. As such, it's trivial (with a quiet job queue) to
re-ip the node.

The patch just relaxes the checks in LUAddNode, and manual testing shows
that live instances on that node keep working, and that cluster verify
is happy afterwards.

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

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

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

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

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

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

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

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

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

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

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

This will be used by the new import/export daemon to split
output by its child processes into lines.

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

There are a few issues with disabled file storage:
- cluster initialization is broken by default, as it uses the 'no'
  setting which is not a valid path
- some other parts of the code require the file storage dir to be a
  valid path; we workaround by skipping such code paths when it is
  disabled

A side effect is that we abstract the storage type checks into a
separate function and add validation in RepairNodeStorage (previously a
luxi client which didn't use cli.py and submitted an invalid type would
get "storage units of type 'foo' can not be repaired").

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

Commit 43575108 added bootstrap.GenerateclusterCrypto and commit
7506a7f1 changed cfgupgrade to use it. However, this lost the
functionality of upgrading in non-default DATA_DIR.

To fix this, we enhance bootstrap.GenerateclusterCrypto to accept custom
file paths for the three files it modifies. If more files will be needed
in the future, we could just pass in modified DATA_DIR, but for now it
suffices.

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

* Be more clear about when POST and PUT should be used
* Be more verbose on parameters
* Describe allowed parameter values (if not self explaining)

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

When importing an instance, all the saved valued will be used as
explicitly specified values, overriding the cluster defaults. This means
export+import will change the status (from default to explicitly
specified) of parameters.

This patch adds a new option that changes the behaviour to identify
parameter values which are equal to the current cluster defaults and
mark them as such. It does this for hv, be and nic parameters.

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

Currently the instance creation checks the cluster hv defaults + the new
parameters for validity, ignoring the os-specific hvparams (this was an
oversight during the implementation of the os hvp). This patch uses the
new GetHVDefaults method to provide the defaults, instead of reading
directly the attributes of the cluster object.

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

Currently the FillHV method is the one that does the cluster hvparams +
os hvparams merger. However, in some cases we need to do just this,
without adding the instance hvparams on top.

This patch adds a function to compute just this (hv + os hvp
combination) default dict, and modifies FillHV to use it to build the
final dict.

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

If the user doesn't pass any nics in import, do not use a default
one-nic, but instead read the nics from the export file as is.

Fortunately the export and the way nics are read from the command line
are compatible…

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

Similar to the previous patches, if we're missing some parameters and
the export has them (either in the new style or old-style), we reuse
them.

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