Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

It is not actually used.

Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

This options file for ghci preseeds the correct include paths, so that
interactive sessions don't need to always pass these args.

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

This patch converts all the call paths from 'Result' (which contains
just string errors) to 'ErrorResult', which holds
GanetiException-encoded errors. We can now return proper
OpPrereq/OpExec errors to the clients of the luxi/query socket.

The patch touches many files as we had to convert the entire call
chains in a single round. But it should be pretty straightforward
otherwise:

- change 'Result' into 'ErrorResult'
- add error annotations: change "Bad msg" into "Bad (XXXEror msg)"
- add a helper function for confd, where we don't send to client
  formatted exceptions, to convert back from ErrorResult into Result
- change tests similarly, where needed

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

In Python, formatError also returns the exit code, but I find that
splitting them leads to clearer code.

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

As described in the module doc string, while writing this it dawned
upon me that we're mixing all errors together into a single hierarchy
(well, type on the Haskell side), which is not good. Some errors are
used purely within noded, some in the CLI frontends, etc. so these
should not be the same type; frontend functions should only be able to
raise frontend errors, not backend ones.

As to this patch itself, I've used again Template Haskell to generate
both the data type and the serialisation functions, as the initial
version, hand-written, seemed too prone to errors due to string
matching.

A small unittest for checking serialisation consistency is also added.

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

These are almost generic, so let's change the signatures a bit a make
them fully so.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Adeodato Simo <dato@google.com>

This 'simple' way of defining objects will be used also for errors, so
let's make it less Luxi-specific.

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

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

The 'dumb-allocator' has been removed almost two years ago (commit
6f547f96, “Remove dumb-allocator”), let's remove this special casing
for it from devel/upload.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

While running with a wrong --prefix/--sysconfdir, I saw that
devel/upload actually uses a hardcoded path for the init script, even
though it installs it in the correct place.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

This documents the status (or wanted status) of some example ocf modules
I've written for Ganeti. They are far from perfect, but they should be
shipped as a starting point for other people who want to run this to
improve on.

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

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

To people not used to it, it was completely non-obvious why
"./devel/upload" didn't do anything.

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

Currently, when we have an assertion error raised from cmdlib, it looks like this:

  [cluster] root@node4:~# gnt-instance grow-disk instance1 0 1G
  Failure: command execution error:

This is very very confusing. This patch adds a bit of traceback
formatting to improve this as follows:

  [cluster] root@node4:~# gnt-instance grow-disk instance1 0 1G
  Failure: command execution error:
  Internal assertion error: please report this as a bug.
  Error message: ''; location:
    File "/usr/lib/python2.6/dist-packages/ganeti/cmdlib.py", line 11954, in CheckPrereq
      assert False

This is not perfect, but at least it shows better what the problem is.

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

Pep8 didn't agree with the indentation.

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

This is a new tool as per the design document “design-ssh-setup”. It
receives a JSON data structure on its standard input and configures the
SSH daemon and root's SSH keys accordingly. Unit tests are included.

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

This patch changes “ssh.GetUserFiles” to support two different kinds of
SSH keys, RSA and DSA. Before it would always use DSA. Newly written
unit tests are included.

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

Signed-off-by: Michele Tartara <mtartara@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

- Debian Squeeze and up have a package named “python-coverage”, but it
  doesn't use the same binary name as upstream (“coverage”).
- Said package includes a patch to use symlinks instead of file copies
  for jQuery. If files from previous runs are around, an exception is
  raised. This is fixed by removing all regular files and symlinks
  before generating a report.

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

This is just a bit of cleanup. The (.&&.) operator is internally just:

  a .&& b = conjoin [a, b]

so let's replace 'a .&&. b .&&. c .&&. d' directly with 'conjoin [a,
b, c, d]'.

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

After seeing how nice HUnit formats the error message on failed
'assertEqual', I think we can do better with ==?. Currently it says
(on one line): "Expected equality, but 1 /= 2".

This patch changes the code to format it similar to HUnit:

  Expected equality, but got mismatch
  expected: 1
   but got: 2

(on three lines). This makes it more clear what is the expected and
what is the wrong value.

A few tests have been modified to ensure that the expected value is
the second argument to ==?. This is different than HUnit, but makes
more sense in the operator version (I think).

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

This is used when one wants to build all binaries, including those
that are used only for testing. A handy shortcut to make sure all
binaries can be built.

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

This includes:
 * The operator (/=?), which checks for inequality and prints
   an error message if it encounters equality. (Basically the
   negation of the (==?) operator).
 * Application of this operator in the test property
   prop_addPri_NoN1Fail.

Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This patch includes:

* The 'failN1' flag is now only set if there is strictly less
  memory available than required for failover.
* Unit tests for that.

Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Small simplifications of other unit tests using the (==?)
operator when possible, and typo fixes.

Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

eitherToResult now converts from `Either a b` to `GenericResult`, not
necessarily from `Either String` only. Also, fix a typo.

Signed-off-by: Dato Simó <dato@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This was asked a few times offline, so let's document it.

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

We also need to change the signature for _GetInstanceInfo, since we
don't have access to the instance dictionary in the rename instance
test.

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

Currently, we set "info" metadata on block devices at device creation
time, but we never update it, leading to stale data in case of
instance renames. This would not be a big problem in case of regular
renames (assuming this is a rare operation), but importing instances
into the cluster via the import/export feature usually is done with a
rename; this means that all imported instances have wrong information
in their block devices.

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

This patch adds a small helper function to clear an LV's tags, and
calls it at SetInfo time. We need this to be able to correctly track
instance renames, once we will call SetInfo at such times.

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

I wanted to write that snippet the third time, which is too much :)

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

This can indeed be annoying, so let's document it.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Adeodato Simo <dato@google.com>

While writing the initial version of this design I misunderstood what
“setup-ssh” does. It doesn't actually start the node daemon and just
configures SSH, nothing else.

This patch adds the cluster name as a field to the JSON structure and
clarifies the purpose of the node daemon certificate given.

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

For simple SSH keys, that is those without options such as
“command="…"”, only the first two parts need to be compared. The third
field is a free-form comment.

This patch changes the comparison used in
AddAuthorizedKey/RemoveAuthorizedKey to take this into account. Lines
with options are still compared in full.

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

This can be used to detect a reboot between two checks.

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

There might be more, but at least replace all these low-hanging fruits.

Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Helga Velroyen <helgav@google.com>

This design addresses the lack of a uniform way to query ganeti nodes
for real time information that can be used by monitoring.

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

Since we now have the GeneralResult as a multi-purpose monad, we can
remove the custom OpResult monad, and just use 'GeneralResult
FailMode' as our type. This allows removal of a few bits of
specialised infrastructure, relying instead on the generic one.

The restriction on using OpResult as a general monad remains as
before.

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