GitLab

This exports whether htools was enabled at configure-time, and adds a
constant for our reference iallocator.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

There was a race condition on heavily loaded testsystem causing randomly
to fail the timeout unittests as the signal handler is not yet setup but
the timeout has already hit.

Therefore we introduce a workaround to wait until a program reached a
certain point (for example after signal handling setup) before we
actually go for the real run. The wait of course has a timeout as well,
but it's pretty high. If we hit the 20 seconds we have really big issues
anyway.
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

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

Ganeti 2.1.3, released in June 2010, added support for a new, extensible
instance creation request format, called version 1. This patch removes
support for the old and undocumented version 0 format.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

- Added @rtype and/or @return where missing
- Fixed @param for Query() filter_ parameter (colon was missing)
Signed-off-by: Simeon Miteff <simeon.miteff@gmail.com>
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

This fixes a issue, where an stopped instances is reported as ERROR
in cluster verify if it lives on a offline node. As the instances is
down this shouldn't happen.
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

The design details can be seen in the design document
(doc/design-lu-generated-jobs.rst).
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

Add some debug logging to detail why we don't run some steps.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

I provided unittest to test the important pieces of the infrastructure.
The one remaining function (ResuriveEnsure) is not easy to unittest
but also not critical if it fails to operate correctly.
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

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

- Use constants
- Don't sleep if no delay is given
- Mark function as deprecated: it uses polling instead of waiting for changes
  (but the latter needs authentication); it can still be used
- Add unittests
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

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

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

Commit dd7f6776 added another call to BuildHooksEnv to provide
post-phase status variables. Since BuildHooksEnv also built the node
lists, that meant they have to be built twice. First a rather strict
check was used, but it turned out to be more tricky. Commit b423c513


had to remove the strict check again.

With this patch the function is split in two parts, one generating the
actual environment variables, and another part returning the node lists.
The former is called twice.

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

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

This adds a new method WaitForJobCompletion that can be used for
client who are not interested in the entire job log, just in its
completion status.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

Commit dd7f6776 added a restrictive check for the node lists returned
by BuildHooksEnv, leading to errors with some LUs, one of which was
fixed in commit 0dfa2c22

. As it turns out, other LUs have similar
issues, some not easy to fix. This patch disables the restrictive check
until the BuildHooksEnv functions can be split into one part generating
the actual environment and one generating the node lists.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This broke QA (and everyone trying to add a node) by complaining about
different node lists.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

When a hook is called, it is provided with a number of variables
describing the status of the instance/node/etc. before the operation.
Some opcodes provide extra variables to see modified values from hooks,
but that's not a generic solution.

This patch modifies the code calling hooks to generate the environment
once before and once after an opcode has been executed. Doing so should
be safe—I did not find any LU.BuildHooksEnv modifying LU instance
attributes. The values collected after running the opcode are prefixed
with “GANETI_POST_”, as opposed to “GANETI_” for pre-execution
variables. The latter are still provided for backwards compatibility.

Environment variable examples:

gnt-instance start $instance:
GANETI_INSTANCE_STATUS=down
GANETI_POST_INSTANCE_STATUS=up

gnt-instance modify -B memory=512 $instance:
GANETI_INSTANCE_BE_memory=768
GANETI_POST_INSTANCE_BE_memory=512
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Also replace explicit loop with dict.update.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

- Dictionary indentation
- Add empty lines for readability
- Simplify conditional code
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>

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

- Access is only permitted for authenticated clients (queries can return
  sensitive data)
- Filters can be specified when sending a PUT request
- Updates RAPI client, documentation and tests
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

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

Instead of going via cmdlib and using special cases for different
resources, the list of fields is used directly.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

Commit 28b71a76

 added a list of resources which can be queried using
LUXI. Unfortunately the variable was named “QR_OP_LUXI”, which can be
confusing. This patch renames “QR_OP_QUERY” to “QR_VIA_OP”, “QR_OP_LUXI”
to “QR_VIA_LUXI” and adds a new variable, currently equal to
“QR_VIA_LUXI”, for remote API queries (mostly for documentation).
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

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

Previously we failed later with a rather useless error message. This
patch fixes this and tells the user to activate-disks if replace-disks
is in the need of activated disks rather than abort with a cryptic error
message.
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

The behaviour of LUOsDiagnose needs special treatment. Commit d22dfef7


changed it to not return hidden, blacklisted or invalid OSes if the
respective field is not requested. This behaviour needs to be preserved
for users of the old opcode, but since it provides an implicit
filtering, it should not be done for LUQuery (the generic way to query
for resources) and is therefore moved from the underlying implementation
to LUOsDiagnose. Users of the new opcode should pass an explicit filter.
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's not perfect, but at least some more.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Allows converting the value of a column before returning it. Useful for
sorting while still using one of the other generic functions for
retrieving the value.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>

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

- Remove duplication by merging two MatchNameComponent into a wrapper
- Reduce lock duration by getting list of names under lock and then
  matching names without the lock
- Also, ExpandNodeName's docstring is fixed.
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>

Patches db366d9a and aac4511a

 added support for EXT_MIRROR instances,
but inadvertently introduced a bug: for INT_MIRROR cases, we don't
need (actually we can't support) neither an iallocator nor a target
node.

To fix this, we move the iallocator/node checks in CheckPrereq (or
respectively in the tasklet CheckPrereq), where we have access to the
instance configuration, and additionally we check for and prevent
passing either of these two for INT_MIRROR instances.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: René Nussbaumer <rn@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>

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: Iustin Pop <iustin@google.com>

The command line option --power-delay sets the time waited between power
ons.
Signed-off-by: René Nussbaumer <rn@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>