This patch adds LU and cli-level support for modification of the node
drained flag. It is similar to the offline changes.

Reviewed-by: imsnah

This patch adds checks for drained nodes in the logical units that
allocate or move instances around. We also update an error message (not
style-compliant).

Reviewed-by: imsnah

This patch changes slightly the output of gnt-cluster verify for drained
nodes, and also adds a note with the total number of drained nodes
(similar to the offline nodes note).

Reviewed-by: imsnah

This patch changes the master candidate pool computations in
ConfigWriter to properly handle drained nodes. They are now excluded
from counting towards the reachable number of candidates.

The patch also adds verification of consistency for the node status.

Reviewed-by: imsnah

This patch updates burnin not to use drained nodes (similar to the
handling of offline nodes).

Reviewed-by: imsnah

This patch changes the dumb allocator not to use drained nodes (similar
to offline nodes).

Reviewed-by: imsnah

This patch exports the drained attribute:
  - LUQueryNodes accepts now the drained field
  - RAPI exports it for node objects
  - gnt-node info shows it now (along newly-added master_candidate and
    offline flags)
  - gnt-node list can list it (but not by default)
  - to the iallocator scripts

Reviewed-by: imsnah

This attribute will be used to prevent any allocation on the node (any
of replace-disks with new secondary this node, failover to the node,
migration to the node).

The patch adds the attribute and initializes it correctly in cluster
init and for new nodes.

Reviewed-by: imsnah

Reviewed-by: imsnah

This patch converts the _CheckMetaSize method to raise exceptions
instead of logging and returning False. This fits now in the new rpc
return types, so it's a cheap change.

Reviewed-by: ultrotter

This big patch converts the bdev Assemble() methods and the supporting
functions to raise exceptions instead of returning False. This is a big
patch, since the assembly functions touch other functions: add children,
creation, etc. However, the patch does not add much new code, rather it
reworks existing code.

One of the biggest changes is in the rework of the DRBD8._SlowAssemble()
method (one of the most complicated/ugly ones). Hopefully the new
version is a little bit more readable.

Reviewed-by: ultrotter

Currently, the Remove() methods of block devices return True/False.
This doesn't permit any error detail reporting.

This patch changes the return type to None for success, and raises
BlockDeviceError in case of failure. This permits the details to be
passed up the stack.

The patch also simplifies a little the Remove method of file-based
devices (no stat first, just try unlink).

Reviewed-by: ultrotter

This converts the backend and cmdlib modules to a (status, data)
implementation of the blockdev_remove rpc call. bdev.py is not yet
converted, so we don't actually have error information.

We also fix a bug in _RemoveDisks by not reusing a variable.

Reviewed-by: ultrotter

Currently, the Shutdown() methods of block devices return True/False.
This doesn't permit any error detail reporting.

This patch changes the return type to None for success, and raises
BlockDeviceError in case of failure. This permits the details to be
passed up the stack.

For LVM and file-backed devices, this is a simple change. For DRBD, we
first remove the shutdown of disks in case of network activation
failures (since with static minors the minor is used anyway, we don't
gain anything by clearing it), and the we simply change _ShutdownAll()
to raise an exception.

Reviewed-by: ultrotter

This converts the backend and cmdlib modules to a (status, data)
implementation of the blockdev_shutdown rpc call. bdev.py is not yet
converted, so we don't actually have error information.

We also fix a bug in _ShutdownInstanceDisks by not reusing a variable.

Reviewed-by: ultrotter

This converts the RPC call blockdev_assemble to the new-style result
format. Note that we won't usually have error information, but it's the
first step toward it.

Reviewed-by: ultrotter

Child classes of _R_TAGS must define TAG_LEVEL, but for good style let's
define it also here to at least ensure we don't get a 'Unknown
attribute' exception.

Of course, this also silences a pylint warning.

Reviewed-by: amishchenko

In LUSetInstanceParam we used to save the dict without defaults for the
instance params as hv_inst, but to use the populated one for the
instance (hv_new). Fixing this leads to instances without all the
parameters set.

Reviewed-by: iustinp

The comment is not really true anymore, as we have a lot of parameters
nowadays.

Reviewed-by: iustinp

1) Only instance_name is available
2) There was a missing string parameter

Reviewed-by: iustinp

This was skipped in the previous QA patch.

Reviewed-by: imsnah

In some cases 'mouse' may work better than 'tablet', so we'll handle
both by allowing the user to specify a parameter. By default no mouse is
used.

Reviewed-by: iustinp

With this patch we allow KVM instances to be booted off the network.
The only issue is that this is not compatible with virtio nics, so
we disallow them, when booting from the net.

Reviewed-by: iustinp

When executing the KVM runtime we load the nic type from the runtime
hvparams and use it to specify the nic model type. As for the disk we
translate the DEV_PARAVIRTUAL type to 'virtio'.

Reviewed-by: iustinp

They'll be used to set the nic type when we execute the runtime, since
the nics are processed later. We need to save the hvparams because we
want to use the same one as when we saved the runtime, rather than use
the current instance ones, to avoid applying only some changed
parameters when the runtime is loaded.

Reviewed-by: iustinp

By passing the relevant if= value to the disk we support different disk
types. The only change is that we'll translate "paravirtual" to
"virtio" to keep only one "paravirtualized" value, around ganeti. The
if= value is calculated outside the disks loop, as it's the same for all
disks (as currently ganeti doesn't support per-disk params).

Reviewed-by: iustinp

Copy the message from the KVM one, adding a missing 'the' and a list of
possible values, to help the user in his decision.

Reviewed-by: iustinp

- Add a bunch of NICs and DISKs types
- Specify which one are valid disks and nics for KVM (the new ones
  toghether with some of the old ones)
- Add the default values (paravirtual)
- Allow the disk and nic types as parameters and check their validity

Reviewed-by: iustinp

These are not HVM specific, so have been given an HT generic name.

Reviewed-by: iustinp

The VNC base port has nothing to do with HVM itself, and is general to
VNC itself, so we're removing the HT_HVM prefix to the constant.

Reviewed-by: iustinp

This patch adds a new instance query flag called disk_usage that
retrieves the overall space used by an instance on each of its nodes.
This can be used when balancing the cluster or checking N+1 status.

The flag is also exported in RAPI. Note the flag is currently broken for
file-based instances, as it represents the amount of space in the
cluster volume group.

Reviewed-by: ultrotter

Currently, the names of the functions in backend.py that are actually
RPC procedures and are called from ganeti-noded are not corresponding to
the RPC names. This makes it hard to actually see which functions are
exported and which functions are internal to backend.

This patch renames all blockdevice-related functions in backend.py match
the name of the RPC call (without the ‘call’ or ‘perspective’ prefix).
This should make it easier to grep for a given function called in
cmdlib, without having to open and check in ganet-inoded what backend
function it corresponds to.

The patch also does two minor extra cleanups (rename a variable and
change a logging level).

Reviewed-by: ultrotter

This patch adds two utility functions for raising BlockDeviceError
exceptions and for running functions while ignoring this error. Most of
the manual “raise errors.BlockDeviceError” cases are converted to
_ThrowError, as this makes the code clearer.

We also change most of the DRBD error messages to include the minor
number because with the parallel execution of commands it's not longer
possible to identify the failed DRBD just from the timestamp, and the
minor number can be mapped back to the instance easier.

Reviewed-by: ultrotter

This patch converts the call_blockdev_find - which searches for block
devices and returns their status - to the (status, data) format. We also
modify the backend function name to match the rpc call.

Reviewed-by: ultrotter

This is a hand-picked forward patch of commit 1755 on the 1.2 branch
(hand-picked since the trees diverged too much since then):

    The patch changed the xen hypervisor to compute the number of cpu
    sockets/nodes and enables the command line and the RAPI to show this
    information (for RAPI is enabled by default in node details, for gnt-one
    one can use the new “cnodes” and “csockets” fields).

    Originally-Reviewed-by: ultrotter

For the KVM and fake hypervisors, the patch just exports 1 for both
nodes and sockets. This can be fixed, by looking at the
/sys/devices/system/cpu/cpuN/topology directories, and computing the
actual information, but that should be done in a separate patch.

Reviewed-by: imsnah

This patch fixes the error handling in the add OS to instance function
with regard to invalid OSes. Previously, we didn't handle any such
errors, with the end result that the user would have to look in the node
daemon log.

The patch also renames the name of the function to match the RPC call
name.

Reviewed-by: ultrotter

Currently the return value or errors from the block device Open() method
are ignored. This patch catches any BlockDeviceErrors and returns a
well-formatted result.

Reviewed-by: ultrotter

By using the RemoteFailMsg() or the payload field of RpcResult, we can
simplify a few functions in cmdlib.

Reviewed-by: ultrotter

For results which use the (status, payload) response type, it's easier
to define a ‘payload’ field on the result holding the payload than to
extract it using “data[1]” in the caller code.

Reviewed-by: ultrotter

In lockless queries, it's better if we see the instance in ADMIN_down
rather than ERROR_down during the time it's installed. As such, we
change the LU to only mark the instance 'up' at the time we are ready to
start it.

Reviewed-by: ultrotter