We move all disks anyway.

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: 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: Iustin Pop <iustin@google.com>

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
(cherry picked from commit 022c3a0b)

Since drained nodes could be (partially or fully) broken in iallocator,
we ignore results from these nodes when building the cluster map in
preparation for sending it to the script.

This is a cheap change for the stable branch; ideally we should not
query them at all.

The patch also fixes a typo in iallocator.rst.

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

This implements CONFD_REQ_NODE_PIP_BY_INSTANCE_IP, and maps instance ips
to primary node ips.

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

Rather than composing them every time, we'll keep "ready-made" tuples
for unknown entries and internal errors (which is, yet, unused)

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

Add the following functions to SimpleConfigReader:
GetInstanceByIp, nic ip -> instance name
GetNodePrimaryIp, node name -> node primary ip
GetInstancePrimaryNode, instance name -> primary node

For the first one we precalculate the _ip_to_instance, so we don't have
to loop over all instances for each query.

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

This implements CONFD_REQ_NODE_ROLE_BYNAME.

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

Adding a way to retrieve the three node status flags, all in one go,
from SimpleConfigReader. Non-existing nodes are handled by returning
"None".

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

If we standardize common error codes, clients can make use of them to
react, rather than trying to parse error strings.

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

Apart from being a master, a candidate, offline or drained, a node can
have one more state: none of those. Adding back the normal, nice,
healthy "just" a node! :)

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

During a query execution we currently verify whether we have that query
supported in our dispatch table. Since it's an error not to have it
anyway, we move this verification to the class creation, erroring out
early if such a problem exists.

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

We'll start with a simple ping query, which returns ok if it gets no
query argument, and an error if it gets any. Also non-implemented query
are handled by the base ConfdQuery class, so we don't have to
special-case them in ConfdProcessor anymore.

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: Iustin Pop <iustin@google.com>

The message is an error deriving from some code moving/refactoring
happening before the initial submission

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

confd.server and daemons/ganeti-confd import a few modules they don't
actually use. Clean them up.

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

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

This fixes ganeti-rapi which was broken in commit 7a95a954.

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: Guido Trotter <ultrotter@google.com>

Use variables already defined.
Introduce ganeti CONF_DIR.

Signed-off-by: Luca Bigliardi <shammash@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

Use " " instead of ",".

Signed-off-by: Luca Bigliardi <shammash@google.com>

Map the node hostname with its primary/secondary ganeti IP in ssconf.
In this way a node can discover which are its IPs simply grepping ssconf files,
thus avoiding checks depending on network configuration.

Signed-off-by: Luca Bigliardi <shammash@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>

A ConfdProcessor is the job processor for a confd query.
It's responsible for all the processing, from unpacking it, verifying
it, to calculating the result.

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

It will be used for the confd server code, and for a basic client
library.

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

Adding the various basic daemon constants for confd.

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

These are valid values for the "status" field in objects.ConfdReply

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: Iustin Pop <iustin@google.com>

If for some reason we try to call Reload twice but the config file
hasn't changed, we don't want to load it again. Use stat() on the file
to cheaply avoid that situation.

We also return a boolean value, according to whether we have reloaded
the config or not.

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

Since we use it also to load the config file initially we rename the
Reload function to _Load. Then we re-export it as Reload to clients
(since for them it's a reload).

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

Add an 'empty' logical unit to run hooks after cluster initialization.

Signed-off-by: Luca Bigliardi <shammash@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>

This moves yet another hardcoded value to a hypervisor parameter. I
removed the 64/32 difference as it doesn't seem valid to me - it's more
of a local site config rather than arch config.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
(cherry picked from commit 09ea8710)

For the xen-hvm hypervisor, the KERNEL_PATH parameter is needed but
today is hardcoded to a constants in the xen hypervisor library (argh!).

This patch moves this to a hypervisor constant with the default value
being the current hardcoded path. This will allow cluster/instance
customisation based on the installed xen version.

This should fix Debian bug #528618.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
(cherry picked from commit e2ee1cea)

From time to time we're adding new be or hv parameters. With this patch
missing parameters get set to the default value when loading the cluster
object. This patch version also considers the case when hv/be params
don't exist at all, and fixes a broken unit test triggered in that
case.

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
(cherry picked from commit c1b42c18)

If --no-etc-hosts is passed in at cluster init time we set a new
parameter in the cluster's object to false, and avoid adding nodes to
the hosts file. The UpgradeConfig function is used to set the value to
True, when upgrading from an old configuration version.

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
(cherry picked from commit b86a6bcd)

Add a very basic configuration update mechanism to objects.
An object can define the UpgradeConfig method, which will be called at
init time, and use it to fill in missing defaults in the configuration.
In the future we may want to make it more complex, for example adding
the config version, but for now a basic solution will do.

Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
(cherry picked from commit 560428be)

GetConfigSerialNo and GetClusterSerialNo return respectively the serial
number for the config and for the cluster object.

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

Don't indent the final message.

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