Signed-off-by: Stratos Psomadakis <psomas@grnet.gr>

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

Add SHELL_ENV_INIT and XEN_CONFIG_DIR

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

Add new script vif-ganeti that could be used instead of
default vif-bridge in order NIC customization can be more
flexible.

Extract common code from kvm-ifup and insert it in a new
file net-common that is being sourced by kvm-ifup and
vif-ganeti and located under package lib dir (/usr/lib/ganeti/).

Install vif-ganeti properly with make.

Add a mkdir and ln rule in install-exec-hook so that
the /etc/xen/scripts dir is created and the vif-ganeti
scripts gets properly linked inside it.

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>
Signed-off-by: Thomas Thrainer <thomasth@google.com>

For every NIC write down a file that contains NIC's
info (MAC, mode, link, network details), that can
be sourced by any vif script.  The file location is:
/var/run/ganeti/xen-hypervisor/nic/<domname>/<nicidx>.

This file is created upon cfg file creation and before
starting the instance and can be sourced by an external
script.

Cleanup NIC dir when removing config files. Upon Xen
configuration file removal, remove NIC dir as well.

Add new hv param vif_script that allows execution of
user defined networking script and overrides the one
xend is configured with in xend-config.sxp

Ensure that all run dirs exist for Xen hypervisor

Xen hypervisor uses runtime files to store each NIC's related info.
We need to check if parent dirs exist as well.

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>
Signed-off-by: Thomas Thrainer <thomasth@google.com>

Return -1 and not None for vg_size and vg_free and make
volume_group_name optional in cluster config object in Haskell.

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

This is needed in case an instance must be booted with an exact type of
machine version (due to e.g. outdated drivers). If this hv param
explicitly defined (during instance creation/modification) then the kvm
process will be started having this version for the -M option. Otherwise
the default (derived from _GetDefaultMachineVersion() will be used.

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

This will be used in QA to format network interface parameters.

This is a cherry-pick of master commit eac9b7b8

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

This was fixed in stable-2.6, commit 053c356a

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

Requested in issue 337.

The parameter “bridge” was not documented and is therefore silently
replaced with “master-netdev”. A note is added to “qa-sample.json”
describing how comments work.

This is a cherry-pick of master commit
3601d488

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

Commit ec0652ad (June 2009) removed the option.

This is a cherry-pick of master commit
78453739

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

kvm -M ? returns the supported machines (e.g. pc-1.1).
Add _GetDefaultMachineVersion() function to get the default value.

Upon kvm runtime file creation (this is in _GenerateKVMRuntime() invoked
only in StartInstance()) append this info in kvm_cmd. During
live migration the -incoming kvm process is started based on this file.

In case of different KVM versions between source and target nodes
there is a possibility (e.g. due to a kvm bug) for migration to fail silently.

This patch forces the target node to emulate the same machine version
used by running process. If KVM on target node does not support it
the -incoming kvm process will crash and migration will be aborted.

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

I'm already setting this to a release date of tomorrow, since QA on
the 2.6 branch has been clean.

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

Accidentally stumbled upon this while testing unrelated code on a
machine with ~3K active jobs - the bash completion unittest was
hanging.

Upon investigation, it turns out that bash's ${var//pattern/repl/} is
probably quadratic in the size of input (or worse, even):

  $ touch job-{1..500}
  $ time ( a=$(echo job-*); echo ${a//job-/}| wc -c; )
  1892

  real    0m0.597s
  user    0m0.590s

  $ touch job-{1..1000}
  $ time ( a=$(echo job-*); echo ${a//job-/}| wc -c; )
  3893

  real    0m4.654s
  user    0m4.580s

We can easily fix this if we change to array-based substitution (once
per element):

  $ time ( a=($(echo job-*)); echo ${a[*]/job-/} |wc -c; )
  3893

  real    0m0.028s
  user    0m0.010s

  $ touch job-{1..10000}
  $ time ( a=($(echo job-*)); echo ${a[*]/job-/} |wc -c; )
  48894

  real    0m0.233s
  user    0m0.220s

This means that exactly when the master node is busy processing many
jobs, we could accidentally start consuming lots of CPU in the bash
completion, which is not good.

Note: the code might have problems with filenames containing spaces (I
didn't reset the IFS, etc.), but the original code had the same issue,
I think.

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

During a normal configuration update, done via write to temp file and
rename, this is what confd logs (slightly edited for clarity):

2012-12-14 01:05:53: ganeti-confd INFO Loaded new config, serial 21866
2012-12-14 01:06:18: ganeti-confd INFO File lost, trying to re-establish notifier
2012-12-14 01:06:18: ganeti-confd INFO Loaded new config, serial 21867
2012-12-14 01:07:09: ganeti-confd INFO File lost, trying to re-establish notifier
2012-12-14 01:07:09: ganeti-confd INFO Loaded new config, serial 21868

Since this happens always, we should demote the "File lost" messages
to debug level, to keep the logs more clear.

Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Bernardo Dal Seno <bdalseno@google.com>

Note: I'll add tests for this on the master branch, but not here.

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

The only semantic change is the fix of the spelling of the option
--ipolicy-disk-templates.

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

This fixes Issue 257.

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

Commit 1ce03fb1 (“Add ht-based result checks to opcodes”) introduced
infrastructure for checking opcode results, and subsequent commits
improved the list of opcodes which do declare a result, however this
was not tested for dry-run mode operation.

Furthermore, there's no authoritative list of which opcode/LUs support
dry_run mode at all; currently, this is based/restricted on the list
of CLI options, so… for now we disable the result verification if the
opcode has been executed in dry_run mode.

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

This is in preparation for a 2.6.2 release.

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

When all OS parameters should be unset (“gnt-os modify -H -xen-pvm
foo”), a TypeError was raised. This fixes issue 311.

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

This patch is taken from ganeti mailing list: http://goo.gl/LgukY



Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

In _GenerateDiskTemplate() we make use of params variable in case
of ExtStorage. Hotplug code conflicts here.

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

Conflicts:
	lib/cmdlib.py
	tools/cfgupgrade

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

Remove pci_pool from hotplug_info. Hypervisor decides the pci slot
based on current occupations.

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

In cmdlib we must only use the hypervisor class, and never instantiate
it. As such we have to call GetHypervisorClass instead, to avoid
getting an instance of it.

This fixes Issue 316, because __init__ is not called from masterd
anymore, and thus can't fail on EnsureDirs.

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

The 'command' attribute of the OpOobCommand command is defined with a
default value of None, but its validation requires a member of
constants.OOB_COMMANDS, which doesn't accept None. This result in the
following error when submitting an opcode without the command:

  error type: wrong_input, error details:
  Parameter 'OP_OOB_COMMAND.command' fails validation

I suspect this was simply a mistake, since the commit that introduced
it (65e183af, “opcodes: Add opcode parameter definitions”) did lots of
bulk updates.

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

s/exists/exist/

Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Michele Tartara <mtartara@google.com>

Signed-off-by: Dimitris Aragiorgis <dimara@grnet.gr>

The text of the manpage explains that an index can be prepended to
"remove" but the short help doesn't mention it. Adding it helps making
the syntax clear.

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

Commit 6a1434d7 (“Make migration RPC non-blocking”) changed the API
for reporting migration status, but has a small cosmetic bug: if the
migration status if failure, but the RPC itself to get the status
didn't fail, it shows the following error message:

  Could not migrate instance instance2: None

since it always uses result.fail_msg, irrespective of which part of
the if condition failed.

This patch simply updates the msg if not already set, leading to:

  Could not migrate instance instance2: hypervisor returned failure

Proper error display can be done once the migration status objects can
return failure information as well, beside status.

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

Commit 6a1434d7 (“Make migration RPC non-blocking”) changed from
raising HypervisorErrors to returning MigrationStatus
objects. However, these objects don't have an "info" attribute, so
they can't pass a reason back (which is in itself a bug); but the KVM
hypervisor code attempts to do so, and fails at runtime with:

  Failed to get migration status: 'MigrationStatus' object has no attribute 'info'

instead of the intended:

  Migration failed, aborting: too many broken 'info migrate' answers

For now (on stable-2.6), let's just remove the "info" reason, and
later we can add it back properly once we have a way to correctly
represent migration status failures in the LU.

This fixes issue 297.

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

Currently, the code uses createFile, which has the effect of always
truncating the file. This is bad, as the content of the PID file is
wiped even when we wouldn't be able to lock it!

We switch to openFd (createFile is just a wrapper over that), and we
use an explicit set of flags; defaultFileFlags is already safe
(trunc=False), but I prefer to set it explicitly with our desired
flags.

Note that this bug doesn't manifest in normal usage, as daemon-util
won't try to start the daemon if already running. But if anyone or
anything does call ganeti-confd explicitly, the pid file will be
emptied and the daemon will keep trying to be restarted forever…

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

pylint complained, I fixed it, and unfortunately pushed too early.

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

Once again this will be used by forthcoming RAPI test.

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

A newly added test for RAPI will also verify the returned headers. A
test in ganeti.rapi.client_unittest.py is split into smaller stand-alone
tests.

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