Currently, we check if we have a given ip address (i.e. it's alive on
one of our interfaces) but manually calling TcpPing(source=localhost).
This works, but having it spread all over the code makes it hard to
change the implementation.

The patch abstracts this into a separate utils.OwnIpAddress(addr)
function. We add a rpc call for it, which we use instead of the
(single-use of) call_node_tcp_ping. We leave node_tcp_ping in, as seems
useful and eventually it should be removed in a separate patch.

Reviewed-by: imsnah

Reviewed-by: iustinp

Allow multiple api versions in an OS. This is according to the OS API
changes design doc, by which an OS can support multiple versions of the
Ganeti API and if one is supported by Ganeti it will work. Since up to
version 5 of the API mandates an OS could support only one version, this
change is retrocompatible with it and requires no version bump up.

Reviewed-by: iustinp

When calling node_verify leads to an error _VerifyNodes tries to iterate
over a non-sequence. Catch the error before and avoid this from
happening.

Reviewed-by: iustinp

Reviewed-by: iustinp

This patch adds another implementation of an HTTP server. It's
based on code of Python's BaseHTTPServer, from both version
2.4 and 3k. In the future we can write code to decide whether
we should fork for a request or not. Keep-alive is not supported.

Reviewed-by: iustinp

This mainloop can be used in daemons like ganeti-noded.

Reviewed-by: iustinp

This clarifies the job storage and the reason for choosing it.

Reviewed-by: imsnah

With change 1773 a new status WAITLOCK was introduced if a job/opcode is
waiting for a lock. This change updates the document about the job-queue
accordingly.

Reviewed-by: iustinp

This big patch changes the call model used in internode-rpc from
standalong function calls in the rpc module to via a RpcRunner class,
that holds all the methods. This can be used in the future to enable
smarter processing in the RPC layer itself (some quick examples are not
setting the DiskID from cmdlib code, but only once in each rpc call,
etc.).

There are a few RPC calls that are made outside of the LU code, and
these calls are left as staticmethods, so they can be used without a
class instance (which requires a ConfigWriter instance).

Reviewed-by: imsnah

This patch is a cleanup of the standalone functions in cmdlib. Many of
them too as argument a ConfigWriter instance, but some also took other
parameters from the lu (e.g. proc), and in the future, if we want to
also pass the RpcRunner, we would have to add yet another parameter.

One option is to make all these methods of top-level LogicalUnit class.
I took another approach, and made (almost) all these functions take as
first parameter the lu instance. It's like methods, just not declared
under LogicalUnit.

Reviewed-by: imsnah

Indentation in bootstrap was wrong and some names in cmdlib.py were not
right.

Reviewed-by: imsnah

This check can be done earlier, in ExpandNames, and is needed here for
the hypervisor parameter check.

Reviewed-by: ultrotter

Reviewed-by: imsnah

Reviewed-by: iustinp

Reviewed-by: ultrotter

Reviewed-by: iustinp

Since in 2.0 the user will possibly have more interaction with the
hypervisor names, we sanitize them by removing the version numbers
(the version can be a prerequisite for the ganeti installation, we
shouldn't document it in variable names).

Reviewed-by: schreiberal

The idea is that if the OSes will support multiple version (e.g. both
1.2 and 2.0), then Ganeti should be able to talk to it using version
2.0, but then the script needs to be told nicely what version Ganeti is
using.

Reviewed-by: imsnah

Reviewed-by: iustinp

This (big) patch moves the hypervisor type from the cluster to the
instance level; the cluster attribute remains as the default hypervisor,
and will be renamed accordingly in a next patch. The cluster also gains
the ‘enable_hypervisors’ attribute, and instances can be created with
any of the enabled ones (no provision yet for changing that attribute).

The many many changes in the rpc/backend layer are due to the fact that
all backend code read the hypervisor from the local copy of the config,
and now we have to send it (either in the instance object, or as a
separate parameter) for each function.

The node list by default will list the node free/total memory for the
default hypervisor, a new flag to it should exist to select another
hypervisor. Instance list has a new field, hypervisor, that shows the
instance hypervisor. Cluster verify runs for all enabled hypervisor
types.

The new FIXMEs are related to IAllocator, since now the node
total/free/used memory counts are wrong (we can't reliably compute the
free memory).

Reviewed-by: imsnah

This patch changes formatting and the DRBD shared secret details, and
adds master daemon socket details to the security doc.

Reviewed-by: imsnah

Reviewed-by: imsnah

Reviewed-by: imsnah

This is just some additions of not-yet-mentioned docs.

Reviewed-by: ultrotter

Merged r1777 from branches/ganeti/ganeti-1.2

Reviewed-by: imsnah

Reviewed-by: imsnah

Currently the call_instance_migrate call only passes the instance name;
we need to pass the whole object for the hypervisor_type changes (all
the other individual instance rpc calls already pass the instance
object).

Reviewed-by: imsnah

Reviewed-by: iustinp

Reviewed-by: ultrotter

Background: when we have multiple jobs in the queue (more than just a
few), many of the jobs (up to the number of threads) will be in state
'running', although many of them could be actually blocked, waiting for
some locks. This is not good, as one cannot easily see what is
happening.

The patch extends the opcode/job possible statuses with another one,
waiting, which shows that the LU is in the acquire locks phase. The
mechanism for doing so is simple, we initialize (in the job queue) the
opcode with OP_STATUS_WAITLOCK, and when the processor is ready to give
control to the LU's Exec, it will call a notifier back into the
_JobQueueWorker that sets the opcode status to OP_STATUS_RUNNING (with
the proper queue locking). Because this mechanism does not save the job,
all opcodes on disk will be in status WAITLOCK and not RUNNING anymore,
so we also change the load sequence to consider WAITLOCK as RUNNING.

With the patch applied, creating in parallel (via burnin) five instances
on a five node cluster shows that only two are executing, while three
are waiting for locks.

Reviewed-by: imsnah

Reviewed-by: imsnah

Reviewed-by: imsnah

This patch adds a new luxi call that implements auto-archiving of jobs
older than a certain age (or -1 for all completed jobs), and the gnt-job
command that makes use of this (with 'all' for -1).

Reviewed-by: imsnah

This function will be used for auto-archiving jobs via the command line.
The function is pretty simple, we only support up to weeks since months
and higher are not 'precise' entities, and dealing with them would
require us to start using calendar functions.

Reviewed-by: imsnah

Currently there are three function in backend that need the cluster name
in order to instantiate an SshRunner. The patch changes these to get the
cluster name from the master in the rpc call; once the multi-hypervisor
change is implemented, then very few places in which we need the SCR
remain in the backend.

Reviewed-by: killerfoxi, imsnah

Since the objects read from the config file are passed to the various
threads, it's unsafe to re-read the config file (and throw away
ConfigWriter._config_data). As such, we disable the re-reading of the
file (since now the master is the owner the file, it makes not sense to
re-read it), and any modifications to the file must be done offline,
otherwise they will be overwritten.

Reviewed-by: imsnah

Reviewed-by: iustinp

This patch introduces a simple framework for executing jobs in parallel
in burnin (the ExecJobSet function) and the "--parallel" command line
flag.

The patch also changes the instance creation to run in parallel when the
above flag is given. Error handling/instance removal is currently flacky
with this options if there are errors in the instance creation.

We also modify burnin to reuse a single client.

Reviewed-by: imsnah

In case the job object doesn't have a timestamp (which is a separate
issue), the listing should not break. We fix this by changing the
FormatTimstamp function itself to return '?' in case the timestamp
doesn't look good (note that it still can break if non-integers are
returned, but this is unlikely).

Reviewed-by: imsnah