Iustin Pop [Tue, 22 Sep 2009 13:54:20 +0000 (15:54 +0200)]
Change config upgrade to be explicit
Currently the config upgrade is done at each object instantiation, that
means that ganeti-noded will run UpgradeConfig on all objects received
remotely (instances, disks, nics). This is not so good, so this patch
changes it so that only the ConfigWriter runs this method at
configuration load time.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Luca Bigliardi [Tue, 22 Sep 2009 14:24:30 +0000 (15:24 +0100)]
Confd: add instances IPs query
Extend confd to answer queries about instances IPs.
Signed-off-by: Luca Bigliardi <shammash@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Iustin Pop [Tue, 22 Sep 2009 13:02:41 +0000 (15:02 +0200)]
Fix utils.MatchNameComponent for full matches
While ‘test1’ matches both ‘test1’ and ‘test1.example’, it has a full,
exact match and we should return it if that is the case.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Mon, 21 Sep 2009 14:14:38 +0000 (16:14 +0200)]
Design doc changes for UUID tracking of objects
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Mon, 21 Sep 2009 14:11:25 +0000 (15:11 +0100)]
Fix _RemoveDisk for file based instances
During
621b7678 two typos were introduced which prevent file based
instances removal to work correctly. Fixing both of them to what they
were meant to be.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Mon, 21 Sep 2009 13:23:12 +0000 (15:23 +0200)]
Re-wrap locking changes design to 76 chars per line
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Mon, 21 Sep 2009 12:54:20 +0000 (14:54 +0200)]
Revert "Include NEWS file in documentation"
This reverts commit
350ecfecca136f50877898774c6b40014b5c2b1c. It broke
distcheck.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Guido Trotter [Mon, 21 Sep 2009 10:59:21 +0000 (11:59 +0100)]
gnt-instance info: fix nic display
In branch-2.1 the conversion of NICs to Tuples has been unified to the
_NICListToTuple function, which is used by both hook generation and
instance info generation. Unfortunately the ip and mac fields got
swapped from what instance info expected, so the mac is shown as ip, and
vice versa. Fixing this.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Guido Trotter [Mon, 21 Sep 2009 10:13:43 +0000 (11:13 +0100)]
cmdlib._CreateDisks fix a broken result.Raise
The format string has the ": %s" at the end, but no argument is passed,
which of course raises a TypeError. Removing ": %s" as it's added by
the RpcResult Raise() method anyway.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Fri, 18 Sep 2009 15:24:27 +0000 (17:24 +0200)]
Include NEWS file in documentation
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Fri, 18 Sep 2009 15:22:18 +0000 (17:22 +0200)]
Convert NEWS file to RST
- Add markup for titles
- Convert quotes
- Reindent and rewrap to 72 chars per line
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Iustin Pop [Fri, 18 Sep 2009 12:48:22 +0000 (14:48 +0200)]
Unify the instance creation code
Currently the AddInstance in gnt-instance and ImportInstance in
gnt-backup duplicate all of their code except the actual opcode creation
(the parameters to it). By moving this to cli.py (not optimal location,
but we don't have another one), we can use a single copy of the code,
and simply change based on the mode parameter the few different
opcode parameters.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Fri, 18 Sep 2009 11:45:23 +0000 (13:45 +0200)]
Unify multi-instance handling in gnt-instance
Start, stop and reboot have the same code for dealing with
multi-instance handling. This patch moves all that into a single generic
function, and leaves only the building of the specific opcode for the
operation in the individual functions.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Fri, 18 Sep 2009 11:01:03 +0000 (13:01 +0200)]
Fix gnt-instance start/stop/… args
The gnt-instance commands that have the multi-filter can take also zero
arguments, if the --all flag is passed. But they currently are declared
with min=1, and this is a regression due to the recent bash completion
changes.
There is also the problem that --primary/--secondary expect a node and
not an instance, but currently the completion doesn't support this
distinction.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 13:25:55 +0000 (15:25 +0200)]
Remove explicit DEBUG_OPT and add it by default
Since >90% of the commands take the “--debug” option, and all should
actually take it (the gnt-job command is currently missing it), it makes
sense to simply remove this and add it by default in cli.py.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 13:13:25 +0000 (15:13 +0200)]
Change indentation style in commands definitions
Currently, the indentation level for a command's definition depends on
the command name length, and this makes the commands dict unpleasant to
the eye. This patch changes it so that a new line begins right after the
left parenthesis starting the command tuple.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:58:53 +0000 (14:58 +0200)]
Make ganeti-watcher use the standard debug option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:51:58 +0000 (14:51 +0200)]
Move the “--no-shutdown” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:49:42 +0000 (14:49 +0200)]
Move the “--ignore-secondaries” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:48:43 +0000 (14:48 +0200)]
Move the “--reboot-type” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:46:13 +0000 (14:46 +0200)]
Move cluster verify options to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:41:51 +0000 (14:41 +0200)]
Move more cluster init opts to cli.py
Some of these should be added to cluster modify too.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:37:06 +0000 (14:37 +0200)]
Move the “--mac-prefix” option to cli.py
The default has also been changed to None as this will be useful once
cluster modify will be able to change this setting too.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:32:30 +0000 (14:32 +0200)]
Move the “--no-voting” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:30:41 +0000 (14:30 +0200)]
Move the “--yes-do-it” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:24:26 +0000 (14:24 +0200)]
Unify the “--vg-name” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:21:33 +0000 (14:21 +0200)]
Unify the “--candidate-pool-size” option
This required a slight tweak since in init we want a different default.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:17:42 +0000 (14:17 +0200)]
Unify the “--nic-parameters” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 12:13:41 +0000 (14:13 +0200)]
Unify the “--enabled-hypervisors” option
This also required a change, as one of the cases used a different
default.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:57:50 +0000 (13:57 +0200)]
Unify the “--no-lvm-storage” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:54:19 +0000 (13:54 +0200)]
Move the “--allocatable” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:50:20 +0000 (13:50 +0200)]
Move the node flags options to cli.py
This also unifies a little the parameters to the respective options.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:46:54 +0000 (13:46 +0200)]
Move the “--no-ssh-key-check” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:45:58 +0000 (13:45 +0200)]
Move the “--readd” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:45:21 +0000 (13:45 +0200)]
Unify the “--secondary-ip” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:43:39 +0000 (13:43 +0200)]
Move the “--src-…” options to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:41:31 +0000 (13:41 +0200)]
Move the “--ignore-size” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:40:03 +0000 (13:40 +0200)]
Move the “--auto” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:39:22 +0000 (13:39 +0200)]
Move the “--on-secondary” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:38:35 +0000 (13:38 +0200)]
Move the “--on-primary” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:36:18 +0000 (13:36 +0200)]
Unify the “--new-secondary” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:33:36 +0000 (13:33 +0200)]
Move the “--ignore-failures” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:31:16 +0000 (13:31 +0200)]
Move the “--select-os” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:30:18 +0000 (13:30 +0200)]
Move the “--all” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:29:20 +0000 (13:29 +0200)]
Move the “--static” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:28:16 +0000 (13:28 +0200)]
Move the “--cleanup” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:27:00 +0000 (13:27 +0200)]
Move the “--show-cmd” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:25:48 +0000 (13:25 +0200)]
Move the “--no-start” option to cli.py
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:20:22 +0000 (13:20 +0200)]
Unify the last “--node” (single node) option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 11:07:35 +0000 (13:07 +0200)]
Unify the “--node” (as list of nodes) option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:57:41 +0000 (12:57 +0200)]
Unify the “--node” (as instance placement) option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:54:03 +0000 (12:54 +0200)]
Unify the “--non-live” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:51:12 +0000 (12:51 +0200)]
Unify the “--ignore-consistency” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:48:00 +0000 (12:48 +0200)]
Unify the “--os-size” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:45:08 +0000 (12:45 +0200)]
Unify the “--disks” (disk indices) option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:41:44 +0000 (12:41 +0200)]
Unify the “--disk” (identkeyval) option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:36:54 +0000 (12:36 +0200)]
Unify the “--net” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:33:40 +0000 (12:33 +0200)]
Unify the “--no-ip-check” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:28:24 +0000 (12:28 +0200)]
Unify the last “--hypervisor” (as list) option
This third form takes a list of name+options, and is used only in
gnt-cluster. Who knew that we have three forms of ‘-H’?
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 10:26:28 +0000 (12:26 +0200)]
Unify the “--hypervisor” (with name) option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Thu, 17 Sep 2009 09:51:56 +0000 (11:51 +0200)]
Unify the “--hypervisor” (params only) option
This unifies the two uses of --hypervisor-parameters with only options
(a=b,…) form. The patch changes a little SetInstanceParms since that
expected opts.hypervisor instead of opts.hvparams.
Note that there is another form of --hypervisor-parameters which also
takes the name, that comes in a separate patch.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Iustin Pop [Thu, 17 Sep 2009 09:21:31 +0000 (11:21 +0200)]
Unify the “--backend-parameters” option
This is not straight-forward. First, gnt-cluster had
“--backend-parameters” wheres the rest of the script had just
“--backend”. I settled on the full form, since one can always abbreviate
options, but not vice-versa.
Second, the help in some cases was different, e.g. start instance
had “Temporary backend parameters” and modify had “Change backend
parameters”. I think this kind of distinctions are better served in the
man page. Alternatively, if we could copy options easily, one could
re-instantiate them with a changed help line.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Iustin Pop [Thu, 17 Sep 2009 09:11:22 +0000 (11:11 +0200)]
Unify the “--os-type” option
This was already unified inside gnt-instance, I just moved that
definition (and upper-cased it) into cli.py.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Iustin Pop [Thu, 17 Sep 2009 09:07:23 +0000 (11:07 +0200)]
Unify the “--iallocator” option
This patch also changes the gnt-debug's “--algorithm” into
IALLOCATOR_OPT, since it was the only non-standard use of this (due to
historical reasons). Since it's just a debug option,
backwards-compatibility is not critical here.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Iustin Pop [Thu, 17 Sep 2009 09:04:00 +0000 (11:04 +0200)]
Unify the “--file-driver” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Iustin Pop [Thu, 17 Sep 2009 09:02:07 +0000 (11:02 +0200)]
Unify the “--file-storage-dir” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Iustin Pop [Thu, 17 Sep 2009 08:59:55 +0000 (10:59 +0200)]
Unify the “--no-nics” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Iustin Pop [Thu, 17 Sep 2009 08:57:46 +0000 (10:57 +0200)]
Unify the “--disk-template” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Iustin Pop [Thu, 17 Sep 2009 08:57:03 +0000 (10:57 +0200)]
Unify the “--no-wait-for-sync” option
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Thu, 17 Sep 2009 20:58:23 +0000 (21:58 +0100)]
Explain the reason for fqdn hostnames
The install guide asks the user to use the fqdn in the hostname, but
doesn't explain why. This has been asked multiple times, so adding an
explanation there.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Reviewed-by: Olivier Tharan <olive@google.com>
Michael Hanselmann [Wed, 16 Sep 2009 16:16:05 +0000 (18:16 +0200)]
Remove RpcResult.RemoteFailMsg completely
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Wed, 16 Sep 2009 16:10:12 +0000 (18:10 +0200)]
Add locking improvements to Ganeti 2.1 design document
Also increase the table of contents' depth to 4.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Iustin Pop [Thu, 17 Sep 2009 08:33:48 +0000 (10:33 +0200)]
Add an error-simulation mode to cluster verify
One of the issues we have in ganeti is that it's very hard to test the
error-handling paths; QA and burnin only test the OK code-path, since
it's hard to simulate errors.
LUVerifyCluster is special amongst the LUs in the fact that a) it has a
lot of error paths and b) the error paths only log the error, they don't
do any rollback or other similar actions. Thus, it's enough for this LU
to separate the testing of the error condition from the logging of the
error condition.
This patch does this by replacing code blocks of the form:
if x:
log_error()
[y]
into:
log_error_if(x)
[if x:
y
]
After this change, it's simple enough to turn on logging of all errors
by adding a special case inside log_error_if such that if the incoming
opcode has a special ‘debug_simulate_errors’ attribute and it's true, it
will log unconditionally the error.
Surprisingly this also turns into an absolute code reduction, since some
of the if blocks were simplified. The only downside to this patch is
that the various _VerifyX() functions are now stateful (modifying an
attribute on the LU instance) instead of returning a boolean result.
Last note: yes, this discovered some error cases in the logging.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Wed, 16 Sep 2009 09:55:07 +0000 (11:55 +0200)]
Introduce parseable error codes in LUVerifyCluster
Currently the output of cluster verify can be parsed for 'ERROR'
messages, but that is the only indication we get (error or no error). In
order to allow monitoring tools to separate different error conditions,
this patch introduces a new output format (“gnt-cluster verify
--error-codes”) that changes the output from human-friendly to
machine-friendly. In this mode, an error line changes from:
ERROR: node node1: drbd minor 1 of instance inst1.is not active
to:
ERROR:ENODEDRBD:node:node1:drbd minor 1 of instance inst1 is not active
i.e. the error message is a ‘:’-separated field, with ERROR in the first
place, the error code in the second, the object type (cluster, node,
instance) in the third, the name of the object (for nodes/instances) in
the fourth, and then the text message.
The patch also removes some of the verbosity of the operation
(“Verifying instance X”, “Verifying node X”) since on big clusters these
informational messages can quickly fill up an entire screen. The
original behaviour can be restored via the ‘--verbose’ option.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Tue, 15 Sep 2009 09:42:51 +0000 (11:42 +0200)]
Unify the multiple 'verbose' options
Currently multiple files define a 'verbose' option. This patch moves all
these definitions to a single VERBOSE_OPT option which is exported from
cli.py.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Tue, 15 Sep 2009 09:40:02 +0000 (11:40 +0200)]
Rewrite the lib/cli.py export list
Right now, the export list is a mess, it's not sorted, and the exported
items are in random places. This patches does grow the length of the
list, but at least it's more clear.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Michael Hanselmann [Wed, 16 Sep 2009 15:39:03 +0000 (17:39 +0200)]
Show "--help" in bash completion
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Wed, 16 Sep 2009 13:15:22 +0000 (15:15 +0200)]
Remove RpcResult.failed attribute
It's redundant with RpcResult.fail_msg. This change survived QA, burnin and
unittests.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Luca Bigliardi [Wed, 16 Sep 2009 13:19:18 +0000 (14:19 +0100)]
Confd: add primary IPs queries
Extend confd to answer queries about nodes primary IPs and master candidates
primary IPs.
Signed-off-by: Luca Bigliardi <shammash@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Wed, 16 Sep 2009 11:00:03 +0000 (12:00 +0100)]
ganeti-confd: remove partial imports
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Tue, 15 Sep 2009 16:33:23 +0000 (17:33 +0100)]
Confd client library
Initial confd client library implementation. This initial version uses
asyncore, and supports answers via a callback.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Tue, 15 Sep 2009 12:50:19 +0000 (13:50 +0100)]
Move fourcc packing/unpacking to main confd module
This way it can be used by the client as well
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Tue, 15 Sep 2009 11:20:40 +0000 (12:20 +0100)]
Add unittests for confd constants
Checking that the fourcc code is indeed a fourcc code, and that request
codes and reply statuses are distinct.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Mon, 14 Sep 2009 16:13:45 +0000 (17:13 +0100)]
ganeti-confd: explicitely log failed big sends
Make sure that if we try to send packages which are too big (which
shouldn't happen) this gets properly logged in the config file.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Luca Bigliardi <shammash@google.com>
Guido Trotter [Mon, 14 Sep 2009 15:59:03 +0000 (16:59 +0100)]
Add a magic fourcc code to confd packets
This will make it easier to change the protocol later on
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Mon, 14 Sep 2009 15:42:33 +0000 (16:42 +0100)]
AsyncUDPSocket: Move to a well defined UDP size
Currently we read maximum 4K packets, and don't check packets when
sending them. With this patch we move to a well defined maximum size of
60K.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Mon, 14 Sep 2009 14:54:37 +0000 (15:54 +0100)]
ConfdAsyncUDPServer: fix a docstring
It refers to an older input variable
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Mon, 14 Sep 2009 13:13:17 +0000 (14:13 +0100)]
Abstract AsyncUDPSocket to daemon
This allows this extended asyncore+udp module to be used also in other
daemons, and in the confd client library
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Mon, 14 Sep 2009 10:49:49 +0000 (11:49 +0100)]
ConfdAsyncUDPServer: defer handling writes
Currently if we fail writing to the socket (perhaps because a signal was
delivered) we lose the data we were sending. Although this is not too
bad (it's udp, and data may get lost anyway) we try to avoid this by
just putting the packets to write in a queue and handling that through
the normal asyncore methods.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Guido Trotter [Mon, 14 Sep 2009 10:40:36 +0000 (11:40 +0100)]
ConfdAsyncUDPServer: handle signals at read time
Currently if a signal is delivered during an attempted read, an
exception is logged in the logfile. There is no need for this, so we
handle this case explicitely.
Signed-off-by: Guido Trotter <ultrotter@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Luca Bigliardi [Wed, 16 Sep 2009 12:37:09 +0000 (13:37 +0100)]
Fix authorized_keys generation at cluster init
Copy pub_key in authorized_keys.
Signed-off-by: Luca Bigliardi <shammash@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Michael Hanselmann [Tue, 15 Sep 2009 12:14:23 +0000 (14:14 +0200)]
Move developer notes into normal documentation
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Michael Hanselmann [Mon, 14 Sep 2009 15:51:25 +0000 (17:51 +0200)]
jqueue: Remove unused run_op_index attribute
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Tue, 15 Sep 2009 11:09:31 +0000 (13:09 +0200)]
Export new lock_status field to gnt-job
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Tue, 15 Sep 2009 11:00:41 +0000 (13:00 +0200)]
Keep lock status with every job
This can be useful for debugging locking problems.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Michael Hanselmann [Tue, 15 Sep 2009 11:02:56 +0000 (13:02 +0200)]
Move OpCode processor callbacks into separate class
There are two major arguments for this:
- There will be more callbacks (e.g. for lock debugging) and extending the
parameter list is a lot of work.
- In the jqueue module this allows us to keep per-job or per-opcode variables in
a separate class. Instead of having to clean up the worker class after
processing one job, these references will automatically go out of scope.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Iustin Pop <iustin@google.com>
Iustin Pop [Mon, 14 Sep 2009 11:32:54 +0000 (13:32 +0200)]
Replace last hardcoded 'nodelist' with NV_NODELIST
LUAddNode has this usage remaining, as it's the only other LU to use
call_node_verify (beside LUVerifyCluster).
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Iustin Pop [Mon, 14 Sep 2009 11:25:18 +0000 (13:25 +0200)]
Treat virtual LVs as inexistent
Currently, “gnt-cluster verify” and “gnt-cluster verify-disks” use the
list of LVs as returned by backend.GetVolumeList to determine whether an
LV exists or not. However, LVs can also be ‘virtual’, which is handled
correctly (i.e. as missing) by the bdev code, but not by this function.
This patch changed GetVolumeList to simply skip virtual LVs; this makes
cluster verify and verify-disks report these correctly as missing. The
only downside is that an user could get confused (lvs reports the volume
as existing, but ganeti as missing). However, this is better than simply
considering virtual LVs as “good”.
No other code beside these two gnt-cluster operations uses the
GetVolumeList function, so we don't change the behaviour of the rest of
the code (e.g. replace-disks, instance info, etc.).
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
Michael Hanselmann [Mon, 14 Sep 2009 11:05:21 +0000 (13:05 +0200)]
locking: Acquire SharedLock in shared mode in separate function
This is for some more symetry with
SharedLock.__exclusive_acquire.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Guido Trotter <ultrotter@google.com>
Michael Hanselmann [Wed, 9 Sep 2009 15:51:35 +0000 (17:51 +0200)]
RpcResult: Fix cases where “failed” wouldn't be set to True
This broke “gnt-instance replace-disks --auto” when the instance
is down.
Signed-off-by: Michael Hanselmann <hansmi@google.com>
Reviewed-by: Olivier Tharan <olive@google.com>