GitLab

The list-drbd QA test assumed falsely that no instance has more than
one disk. This can happen, and this patch accounts for it.
Signed-off-by: Hrvoje Ribicic <riba@google.com>
Reviewed-by: Klaus Aehlig <aehlig@google.com>

The UUID/name switch avoided this particular bit of code, and as a
result the list-drbd command failed as it tried to compare UUIDs and
names. This patch fixes the retrieval, converts the newly returned
UUIDs to names, and modifies the QA to the results and not only the
invocation are checked.
Signed-off-by: Gerard Oskamp <gjo@google.com>
Signed-off-by: Hrvoje Ribicic <riba@google.com>
Reviewed-by: Petr Pudlak <pudlak@google.com>

For consistency, this patch renames 'VALID_STORAGE_TYPES'
to 'STORAGE_TYPES', because the set of valid disk
templates is also just names 'DISK_TEMPLATES'.
Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

This improves the QA code wrt to file storage in the
following ways:
- It fixes a FIXME to make the file storage directory
  customizable in the QA config. Fallback is the default
  value.
- It makes QA derive the enabled storage types from
  the enabled disk templates and runs the 'gnt-node
  storage tests' only on the enabled ones.
Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Klaus Aehlig <aehlig@google.com>

This patch fixes the node storage QA, which checked whether
file storage is enabled by looking at the constant
'ENABLE_FILE_STORAGE'. QA should rather retrieve this
information from the QA config.
Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Klaus Aehlig <aehlig@google.com>

If the node count in a cluster exceeds the master candidate pool size
(which is 10 by default), it's no longer possible to remove the master
candidate role from a node without passing --auto-promote (even with
--force). This behavior is perfectly fine, but the TestModifyNode QA
test didn't account for this.

TestModifyNode is changed such that it first removes the master
candidate role from the tested node with --auto-promote. This will, in
big clusters, promote an additional node to a master candidate which
ensures that the following tests succeed.
Signed-off-by: Thomas Thrainer <thomasth@google.com>
Reviewed-by: Klaus Aehlig <aehlig@google.com>

Make it possible to optionally run gnt-node migrate during QA.
Signed-off-by: Thomas Thrainer <thomasth@google.com>
Reviewed-by: Klaus Aehlig <aehlig@google.com>

This patch adds the currently known and used storage methods to the
constants.py file. Note that storage types are different from disk
templates. For example both disk templates, drbd and plain, use lvm
storage types.

The rapi documentation asserts that valid storage types are limited
to lvm and file. With the extension of storage types, this is not true
anymore, but the implementation of listing storage units is not
implemented for the new types. We disable the assert for now, adding
a fixme to extend it once the listing of storage units is extended
for the new types as well.

'gnt-node list-storage' currently does not support any other storage
type than 'lvm' and 'file'. We limit the QA to storage types that are
currently implemented for this operation.
Signed-off-by: Helga Velroyen <helgav@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

Currently QA tries to use “gnt-node list-storage” with all storage
types, but when file storage is disabled this will fail.

The patch changes it so that if file storage is disabled, we test that
indeed fails, and otherwise we include it in the regular tests.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michele Tartara <mtartara@google.com>

The new objects support attributes, which are cleaner than
dictionary-style access. Item access (like a dictionary) is going away
in a couple of patches. Also, pylint is better at checking attributes
than dictionary entries.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Instead of having a module-level function in “qa_config” to release a
node, a new method on node objects is used.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

There's never been support for a configuration where nodes in the same node
group have different values of the exclusive_storage flag. This patch
disables the possibility to change the flag for individual nodes.
Signed-off-by: Bernardo Dal Seno <bdalseno@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

Up until now nodes were stored as a dictionary. The keys were hardcoded
in a lot of places and entries modified directly.

This patch introduces a new class for nodes in QA named “_QaNode”. It
still supports accessing details via dictionary syntax, but that will be
removed after a couple of patches changing all users. Unit tests for
“qa_config.AcquireNode” are also included.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

This will be run only if 'haskell-confd' QA configuration has been
enabled.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Agata Murawska <agatamurawska@google.com>

Added test checking that it is possible to remove an instance with drbd
disk when non-primary node is offline.
Signed-off-by: Agata Murawska <agatamurawska@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Andrea Spadaccini <spadaccio@google.com>

- Makefile.am: added QA directory to the paths checked by pep8
- qa/: fixed the reported errors
- Makefile.am: also, added qa_group.py to qa_scripts
Signed-off-by: Andrea Spadaccini <spadaccio@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

s/'/"/ in (hopefully) the right places.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

As per discussion on the man-page[1] update, this functionality should be
removed and replaced by just give the command to run if the user insists
of power cycle/power off the master and refuse to operate.

[1] http://groups.google.com/group/ganeti-devel/browse_thread/thread/95d4879a747cc295



This also removes the --all flag from gnt-node power because it was
buggy and would have added too much logic for special cases.
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This also has the sideeffect of presenting a table with the result
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

LVM PV storage units would always show as allocatable, even when they
weren't. For some reason I have not been able to determine, the function
parsing the attributes (“_GetAllocatable”) was not even called and the
list opcode simply returned the attribute string as the value (e.g.
“a-”).  Removing “@staticmethod” did the trick and then I just moved it
to module level.

A QA test is included.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Right now, the QA code is not covered by pylint, and this shows at
least one low-impact bug.

This patch does the necessary changes to make QA pylint-clean, and the
changes the makefile to run pylint for it.

Notable changes:

- qa_utils.GenericQueryTest: randfields was not used at all, and my
  belief is that it was indented to be used in order not to modify the
  input list; so I replaced randfields with fields, so we only shuffle
  the our local copy
- qa_node.TestOutOfBand was using it's own copy of AcquireNode(), so I
  replaced it with the existing version
- qa_os: was using 'dir' in a couple of places, replaced with dirname
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

On cluster level there's no 'default' because it's the highest cascading
level. Due to this 'default' is a valid value and doesn't mean to remove
the value from the dict like it does on the other levels. To overcome
this we just empty the string.
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

- Handle situations with no non-master node
- Expand node name to make test work when configuration just has short
  names (e.g. “node1” instead of “node1.example.com”)
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

As we can't test this on master anymore (if we flag the node offline we
would change master role on master) we use the first non master node we
find in the configuration
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

- Query all known fields
- Random combinations (using a PRNG with a fixed seed) of fields
- Order of result names
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Currently, 95% of the QA commands are executed in the same way: on the
master, based on a command list and with expectancies for succes:

    AssertEqual(StartSSH(master['primary'],
                         utils.ShellQuoteArgs(cmd)).wait(), 0)

The rest 5% are variations on this theme (maybe the command needs to
fail, or the node is different, etc.). Based on this, we can simplify
the code significantly if we abstract the common theme into a new
AssertCommand() function. This saves ~250 lines of code in the QA suite,
around 8% of the entire QA code size.

Additionally, the output was very cryptic before (the famous "QA error:
1 != 0" messages), whereas now we show a clear error message (node,
command, exit code and failure mode).

The patch replaces single quotes with double quotes in all the parts of
the code that I touch; let me know if that's not OK…
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

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

This reverts commit dc092ece

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

My last setup-ssh patch series removed this option. Though, this is
just a temporary fix as a patch series is coming which revert that
behaviour to the old state
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Guido Trotter <ultrotter@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>

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

Before the arguments weren't valid.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Reviewed-by: ultrotter

To my knowledge they're used nowhere and it's at least slightly
confusing to people adding new QA checks.

Reviewed-by: ultrotter

qa_node.py: Fix typo in message
cmdlib.py: Don't add readded node to node list
ganeti-qa.py: Make sure readd isn't done for master node

Reviewed-by: iustinp