Adding a locking.py file for the ganeti locking library. Its first component is
the implementation of a non-recursive blocking shared lock complete with a
testing library.

Reviewed-by: imsnah, iustinp

This can be used for testing purposes.

Reviewed-by: ultrotter,imsnah

This patch is a first step in reducing the chance of causing DRBD
activation failures when the primary node has not-perfect data.

This issue is more seen with DRBD8, which has an 'outdate' state (in
which it can get more often). But it can (and before this patch, usually
will) happen with both 7 and 8 in the case the primary has data to sync.

The error comes from the fact that, before this patch, we activate the
primary DRBD device and immediately (i.e. as soon as we can run another
shell command) we try to make it primary. This might fail - since the
primary knows it has some data to catch up to - but we ignored this
error condition. The failure was visible later, in either md failing to
activate over a read-only storage or by instance failing to start.

The patch has two parts: one affecting bdev.py, which changes failures
in BlockDev.Open() from returning False to raising
errors.BlockDeviceError; noone (except a generic method inside bdev.py)
checked this return value and we logged it but the master didn't know
about it; now all classes raise errors from Open if they have a failure.

The other part, affecting cmdlib.py, changes the activation sequence
from:
  - activate on primary node as primary and secondary as secondary, in
    whatever order a function returns the nodes
to the following:
  - activate all drives as secondaries, on both the primary and the
    secondary nodes of the instance
  - after that, on the primary node, re-activate the device stack as
    primary

This is in order to give the chance to DRBD to connect and make the
handshake. As noted in the comments, this just increases the chances of
a handshake/connect, not fixing entirely the problem. However, it is a
good first step and it passes all tests of starting with stale (either
full or partial) primaries, with both drbd 7 and 8, and also passes a
burnin.

Note that the patch might make the device activation a little bit
slower, but it is a reasonable trade-off.

Reviewed-by: imsnah

Reviewed-by: imsnah

This patch completes the change introduced in r566 (trunk) and r568
(branch-1.2).

Reviewed-by: imsnah

Currently just the bridge and ip address are passed. Add an environment
variable for the mac address.

Reviewed-by: iustinp

Reviewed-by: imsnah

gnt-backup export used to export the ip and mac of each nic, but not which
bridge it was connected to. Adding this information.

Reviewed-by: iustinp

The gnt-node and gnt-instance list commands have a customizable list of
output fields, but the list is not up to date (in the man page) and not
easily understandable from the ‘--help’ output.

This patch updates the man pages and adds the available fields and
default fields in the ‘--help’ output, as part of the description.

Example:
Usage
=====
  gnt-node list

Lists the nodes in the cluster. The available fields are (see the man page for
details): name, pinst_cnt, pinst_list, sinst_cnt, sinst_list, pip, sip,
dtotal, dfree, mtotal, mnode, mfree, bootid. The default field list is (in
order): name, dtotal, dfree, mtotal, mnode, mfree, pinst_cnt, sinst_cnt.

Reviewed-by: imsnah,ultrotter

Reviewed-by: ultrotter

The new QA tests for instance modify uncovered a bug in the modify
initrd operation when setting the initrd to none.

Reviewed-by: imsnah

It was wrongly deleted when converting

if a in dict.keys():
to
if a in dict:

Reviewed-by: imsnah

Passing a new aliases dict to generic main we can easily support aliases for
compatibility reasons or simply useability.

Reviewed-by: iustinp

LVM code sometimes adds an extra separator at the end of the field list.
Make the code strip it if exists.

Reviewed-by: imsnah

Currently, the function backend._GetVGInfo only checks for errors via
the exit code of the 'vgs' command. However, there are other ways of
failure so we need to also check for valid output before parsing.

Furthermore, the checks on the exit code were reported via a 'raise
LVMError', however this exception is not handled anywhere and so the
remote caller will not get reasonable data.

This patch does two main things:
  - change the calling protocol for this function to not raise an error,
    and instead return the same type of argument always (dict) with the
    requested keys but values changed into None; this allows in the
    parent rpc call node_info to have valid memory information but
    "error" value for disk space, if there's an error with disks
  - check the validity of the output so that in case we fail to parse
    it, we don't abort with a backtrace in the node daemon but instead
    return the default result value (containing errors), and log these
    cases in the node daemon log file

We also bump the protocol version to 11.

Reviewed-by: ultrotter

This patch does two things:
  - checks that the result values from call_node_info are valid integer
    values and aborts otherwise
  - skips disk space computation for the DT_DISKLESS case

The most important point of the patch is the verification of results
from the rpc call, as it prepares for a patch that allows failures to be
better reported from the remote node.

Reviewed-by: ultrotter

The checking of a node's free memory (via rpc.call_node_info) is done in
both start instance an failover. This patch abstracts this call,
together with the appropriate error handling, into a separate function
called _CheckNodeFreeMemory.

The patch also has some related changes:
  - the check is done in prereq and not in exec for start instance
  - the redundant check in exec for failover has been removed

Reviewed-by: ultrotter

The function backend.UploadFile still uses "/etc/hosts" directly instead
of the existing constant; this patch fixes this.

Reviewed-by: ultrotter

Currently the fake hypervisor has hardcoded ‘/var/run’ as a base
directory for its store. This patch adds a constant RUN_DIR that is used
for both the fake hypervisor and for BDEV_CACHE_DIR.

Reviewed-by: ultrotter

This is a merge from the 1.2 branch

Reviewed-by: imsnah

This is a merge from the 1.2 branch

Reviewed-by: imsnah

This patch fixes two name typos and a style issue (which makes pylint
complain).

Reviewed-by: ultrotter

Some new paramenters of the CreateInstance opcode are optional (namely
kernel_path, initrd_path and hvm_boot_order) but their absence makes the code
crash. Fix this by initializing them to a default value if they're not present.

Reviewed-by: iustinp

This patch adds support for specifying and changing the boot device order for
HVM instances. The boot device order specification is ignored for non HVM
instances.

Reviewed-by: iustinp

First step to support custom boot device order for HVM instances, add a
location to actually store that information for the instance.

Reviewed-by: iustinp

Reviewed-by: schreiberal

This patch changes the hypervisor to use the new kernel and initrd
instance attributes. They kernel is handled as expected (use the custom
or the default one), but the initrd is treated a little bit special, in
order to keep compatibility with previous behaviour:
  - if customized and not found, raise an error
  - if default and not found, do not use one

Reviewed-by: imsnah

This patch changes the info subcommand to nicely show the kernel and
initrd paths.

Reviewed-by: imsnah

This patch adds support for defining the kernel/initrd at instance
creation time, using the same interface as in instance modify.

Reviewed-by: imsnah

This patch adds support in ‘gnt-instance modify’ to set the kernel and
initrd paths. The user can pass either 'default' or 'none' (none is not
valid for kernel).

Reviewed-by: imsnah

This patch adds two new attributes to the instance object:
  - kernel_path
  - object_path

The code ignores them and the configuration handling is not broken by
this addition.

Reviewed-by: imsnah

This patch improves the ‘gnt-cluster verify-disks’ command by adding
support for detecting broken volume groups and missing logical volume
names.

As such, we don't try anymore to activate disks for instances that are
not likely to succeed anyway, and instead report them.

Reviewed-by: schreiberal

This patch changes the Assemble() method for logical volumes from a noop
to do a `lvchange -ay` on the logical volume; this ensures that if the
logical volume is not active, we are able to activate and use it.

Reviewed-by: imsnah

This patch fixes the double attach operation in bdev.AttachOrAssemble,
which was an indentation mistake in the first place.

Reviewed-by: imsnah

This patch changes the bdev.DRBD8._GetDevInfo to take a string instead
of a minor, separates the `drbdsetup show` invocation into a new
separate method (bdev.DRBD8._GetShowData) and modifies the rest of the
DRBD8 class to make the appropriate calls.

It also adds a unittest script and data files for testing various cases
of device output.

Reviewed-by: imsnah

Allow the MAC address of an instance to be specified optionally during
instance creation and later to be changed via instance modify.

Reviewed-by: iustinp

This patch implements a first proof-of-concept for HVM support in Ganeti.
Due to the nature of this patch, it is intended for test environments
only.

Reviewed-by: iustinp

Wrap error handling around creating the instance config file.

Reviewed-by: iustinp

Abstract the parts of the Xen hypervisor interface shared by both PVM and
HVM into a generic interface, subclass the special case PVM from that.

Prepatch for adding HVM support.


Reviewed-by: iustinp

Currently, RunCmd is written with the assumption that programs will have
a small stderr output, therefore we read the child's stdout (which can
be big, so we don't want to block the child) and then the stderr (which
is small, so it shouldn't block).

However, with the ‘gnt-cluster verify-disks’ command, we ourselves
generate heavy stderr, therefore we break the ganeti-watcher which runs
the verify-disks via utils.RunCmd.

This patch turns the RunCmd command into an poll-based one, which means
any kind of interleaved output by a child on stdout/stderr will be
handled correctly. Of course, since the output is buffered in memory,
there are other ways to break RunCmd(). But at least this should fix the
common case.

Reviewed-by: hansmi