* stable-2.1:
  Bump version to 2.1.0~rc2
  Update NEWS file and release Ganeti 2.0.5
  Security issue: add validation of script names
  Move the hooks file mask into constants.py
  Improve LUQueryNodes for lockless case
  Ship rapi.rst/rapi.html in the dist archive

* stable-2.0:
  Update NEWS file and release Ganeti 2.0.5
  Security issue: add validation of script names
  Move the hooks file mask into constants.py
  Improve LUQueryNodes for lockless case
  Ship rapi.rst/rapi.html in the dist archive

Conflicts:
	Makefile.am    (reverted, not needed)
	NEWS           (simple fix for RST-ification)
	configure.ac   (reverted, not needed)
	lib/backend.py (adapted to new RPC result style)

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

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

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

This patch modifies burnin to accept the --no-name-check option and also
adds --no-ip-check (which was always set to True before).

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

Since gnt-instance start doesn't do any checks on the IP, it doesn't
make much sense to do so in instance create (with start) if the user
expressly passes in ‘--no-ip-check’. Removing this requirement eases the
no-name-check mode (otherwise one would have to create --no-start and
then only start).

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

This patch adds --no-name-check to gnt-instance add and gnt-backup
import. This is opposite to the opcode parameter (name_check) as it is
similar to ip_check and start.

It also adds it to RAPI and gnt-instance batch-create as a parameter in
the input (JSON-formatted) file.

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

This adds a new opcode parameter ‘name_check’ (similar to ip_check) that
is not required to be present (to easy backwards compatibility for
tools).

It also adds a CheckArguments to LUCreateInstance and changes the
workflow related to instance IP checks and NIC initialisation based on
it.

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

The cluster verify checks for fqdn are done via address lookups, and
there we actually use the FQDN. However, for the ssh hostname check
which is done at node add time, we rely on the default of the “hostname”
command. And Debian for example recently changed the default to return
the shortname unless one passes ‘--fqdn’.

This patch is imported from the Debian packaging.

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

This patch unifies the search for external script to always go through
utils.FindFile and implements in that function a restriction on valid
chars in file names and (additionally) that the passed name is the
basename of the final (absolute) name.

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

This will allow reuse of the same mask for multiple validations.

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

In most uses of LUQueryNodes, we don't take a lock. This means that the
instance data is not protected across GetInstanceList and
GetInstanceInfo, and this can lead to instances not existing anymore.

Switching to GetAllInstanceInfo means that we get a single,
semi-consistent snapshot (since instances can still be modified, but
it's much better).

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

This patch adds the 'cache' parameter for KVM; currently this is only
customisable at the hypervisor level, so it's the same for all drives
(except any CDROM image, which gets the default).

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

On some distributions pyinotify is installed in a different way, and the
actual module just contains an internal pyinotify entry, which is the
actual library. On others the main pyinotify module contains the library
itself. We'll try both, in order to be more compatible.

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

Patch 4352bf6d changed RAPI docs to RST from the previous SGML format
(the last file to be converted to RST).

However, it didn't add rapi.rst to the docrst Makefile variable, and as
such rapi.html was neither build nor shipped in the archive. We fix this
and we also distribute doc/rapi-resources.gen since otherwise “make
distcheck” has issues (since this file is built and thus lives in _built
whereas rapi.rst is shipped and lives in topsrcdir).

Note: yes, this means that no 2.0 release until now had a rapi.html file
(in the .tar.gz archive).

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

By allowing also the primary ip field to be fetched directly, we avoid
one more confd lookup, or dns request, to find out which address the
master node lives at.

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

Change the ClusterMasterQuery to allow a query, and if present accept a
list of fields to be returned. Currently only name and ip are accepted.

This feature will be used by NLD to route the cluster ip over the nbma.

Backwards compatibility is preserved.

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

The documentation for file objects' read method states that if the size
is "negative or ommitted", all data is read; thus we can simplify it to
have size=-1 by default and not have the if test.

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

The DRBD driver can ignore dead disks but not dead meta devices (for
which it refuses to configure the minor). To handle this case, we check
that the meta device is readable and if not we ignore it (the same as
when backend._RecursiveAssembleBD didn't find it).

A needed change is the move of some checks and initialisers before this
test (which is before the super().__init__ call), but that should be
fine.

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

The warning will be generated if the clocks diverge by more
than 150 seconds. Due to the way the RPC system works, we
cannot get exact time differences, e.g. if one of the
queried nodes is broken. The comparision is done using a
time window.

Confd queries will fail if the clock on the client and server
are more than 300 seconds from each other. This check helps
keeping at least the nodes of a cluster in sync.

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

This shouldn't happen, but if it does it's better to fail at this level,
rather than create a broken NIC script, which is hard to debug.

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

Since the rsync/ssh calls are done in parallel, they can't read properly a
password or confirmation about keys from stdin. As such, it's better to enable
batch mode so that they fail right away instead of prompting and then timing
out after a long while.

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

DRBD goes into sync mode for a short amount of time after
executing the "resize" command. DRBD 8.x below version
8.0.13 contains a bug whereby calling "resize" in sync
mode fails.

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>

It can't work on older distributions where simplejson
doesn't have indentation support.

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

When not initialising the cluster, consider all nodes are added, so that
multi-node tests (e.g. export, replace) work correctly (if there are
nodes, of course).

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

This patch removes the quotes from CommaJoin and converts most of the
callers (that I could find) to it. Since CommaJoin does str(i) for i in
param, we can remove these, thus simplifying slightly a few calls.

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

Commit 495cfdf0 removed “nic.bridges” from the default
list for bulk instance list RAPI requests.

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

Comparing a number with None is not a good idea:

  >>> (0 < None, 0 > None)
  (False, True)

This patch also adds build-bash-completion to the list
of checked Python scripts and wraps one line of more
than 80 characters.

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

This reverts commit 6915bc28 based on thread on
ganeti-devel.

Conflicts:

	lib/cmdlib.py (due to the error code classification, trivial)

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

By checking for this explicitly, the errors (SSLEAY_RAND_BYTES, “PRNG
not seeded”) will happen in the start-up phase of the daemon and not
only when executing remote procedure calls.

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

Clarify the fact that temporary HV/BE params in instance start override
and do not extend the configured parameters; and change the instance
list headers from HVM_* to * since many of the parameters apply to KVM
too. Also fix a typo in the rapi documention for '/2/nodes'.

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

This should fix an issue I've seen exactly once during testing. It might have
been caused by parallel RPC calls to archive jobs.

[…] ganeti-noded:112 ERROR Error in RPC call […]
 File "/usr/lib/python2.4/site-packages/ganeti/backend.py", line 2365, in JobQueueRename
   utils.RenameFile(old, new, mkdir=True)
 File "/usr/lib/python2.4/site-packages/ganeti/utils.py", line 322, in RenameFile
   os.makedirs(os.path.dirname(new), mkdir_mode)
 File "/usr/lib/python2.4/os.py", line 159, in makedirs
   mkdir(name, mode)
OSError: [Errno 17] File exists: '/var/lib/ganeti/queue/archive/0'

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

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

For an instance with exactly one NIC:

$ gnt-instance modify --net 1:ip=1.2.3.4 inst1
Failure: prerequisites not met for this operation:
error type: wrong_input, error details:
Invalid NIC index 1, valid values are 0 to 1

For an instance with no NIC at all, it fails with “Invalid NIC index 0, valid
values are 0 to 0”. This is fixed by this patch.

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

This was originally implemented in 0ce8f948 and partially
rolled back in 9b65e0d4. Apart from re-adding the check,
this patch does some housekeeping by renaming the “_helper”
function to “_AddIpAddress”.

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

(1) Both the man page and the online help report the link and mode
fields, which are in the code called nic_link and nic_mode.
(2) Add missing fields to the online help.

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

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

This includes a few paragraphs about daemon-util.

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