Greg Swift
5ecfe884b8
Extend wait_for module to support draining of active tcp connections
2014-08-21 16:15:59 -05:00
Matt Martz
315ac22437
Need to check e.errno instead of os.errno. Fixes #8548
2014-08-11 12:26:52 -05:00
James Cammarata
eb80bd1a56
Break out of loop for sockets in wait_for when there's no data to read
2014-07-14 14:14:13 -05:00
James Cammarata
d80f4cee9d
Merge branch 'wait-for-port-search-regex' of https://github.com/sivel/ansible into sivel-wait-for-port-search-regex
2014-07-14 13:16:11 -05:00
James Cammarata
70abc6f6de
Merge pull request #7374 from hfinucane/wait_for_pipes
...
Allow wait_for to wait on non-traditional files
2014-06-20 15:55:10 -05:00
Matt Martz
04cd7c233b
Allow search_regex to be used in wait_for with port
2014-06-02 07:17:01 -05:00
Brian Coca
fedc3a4f85
corrected doc example, non existing fact to the 2 it had 'merged'
2014-05-29 22:44:33 -04:00
Henry Finucane
6efc8008da
Fail fast in the presence of permissions issues
2014-05-13 09:43:47 -07:00
Henry Finucane
278ecb9b55
Allow wait_for to wait on non-traditional files
...
Use os.path.exists to check for file existence, instead of "can we open
this file for reading".
Fixes #6710
2014-05-12 23:21:55 -07:00
James Cammarata
b71a451ed0
Close fds when running commands from the accelerate daemon
...
Fixes #7307
2014-05-07 08:47:38 -05:00
Dag Wieers
3c15d64655
Add an example to show all hostvars using debug
...
This is a useful example to help debug how facts and vars are being collated.
2014-04-21 11:40:53 +02:00
James Cammarata
3ea5d573aa
Acclerate improvements
...
* Added capability to support multiple keys, so clients from different
machines can connect to a single daemon instance
* Any activity on the daemon will cause the timeout to extend, so that the
daemon must be idle for the full number of minutes before it will auto-
shutdown
* Various other small fixes to remove some redundancy
Fixes #5171
2014-03-23 14:45:49 -05:00
James Cammarata
0556c53f78
Execute commands through accelerate with shell when needed
...
This fixes an issue related to the new run_command changes, whereby
Popen fails to run a command when an executable is specified + shell=False
and there are multiple positional arguments (see Popen docs for details)
2014-03-18 12:24:13 -05:00
Michael DeHaan
47aff528b9
Amend wait_for docs to reference version of feature addition.
2014-03-16 17:10:37 -04:00
Michael DeHaan
de57057cae
Merge branch 'devel' of git://github.com/AnderEnder/ansible into devel
2014-03-16 17:10:02 -04:00
James Cammarata
25cf215f54
Create the tempdir in the accelerate module if it doesn't exist
...
Fixes #6047
2014-03-05 22:08:09 -06:00
Michael DeHaan
90655cb7b0
Fix up examples in the assert docs to match the way things work.
2014-03-01 09:34:21 -05:00
Michael DeHaan
c91ceddfac
Add the start of an integration test structure using Ansible playbooks, also added an assert action plugin to make writing those easier.
2014-02-13 18:28:29 -05:00
James Cammarata
9574f89471
Detect remote_user change in accelerate daemon and allow a restart
...
Fixes #5812
2014-01-30 14:40:41 -06:00
jctanner
b411cce6f3
Merge pull request #5784 from geerlingguy/patch-1
...
Update wait_for documentation - remove extra quote
2014-01-29 09:51:48 -08:00
Michael DeHaan
83f43b82ad
ansible -> ansibleworks
2014-01-28 20:29:41 -05:00
Jeff Geerling
40e5f641e4
Update wait_for documentation - remove extra quote
...
The timeout is in seconds (int), and the extra quote at the end could throw someone off if they're copying and pasting the example.
2014-01-27 12:09:54 -06:00
Jim Kleckner
edca1d69cf
Fix documentation example for the fail module
...
The example for the fail module doesn't work:
http://www.ansibleworks.com/docs/modules.html#fail
The current text shows:
- fail: msg="The system may not be provisioned according to the CMDB status."
when: "{{ cmdb_status }} != 'to-be-staged'"
The "when" documentation indicates that the argument is already a Jinja2
expression:
http://www.ansibleworks.com/docs/playbooks_conditionals.html#the-when-statement
Thus, the following is
when: cmdb_status != "to-be-staged"
is preferred even though the following could work but generates a
deprecation warning:
when: {{cmdb_status != "to-be-staged"}}
2013-12-26 16:32:32 -08:00
Andrii Radyk
f71e67087f
Adding absent condition for to the wait_for module
2013-12-12 10:28:01 +02:00
James Tanner
ad837709bd
Addresses #5023 Fix import comments
2013-12-02 15:13:49 -05:00
James Tanner
6000d636b3
Fixes #5023 Convert all modules to use new snippet import pattern
2013-12-02 15:11:23 -05:00
James Cammarata
30948ae0dc
Fix accelerate module parameter for IPv6 activation
2013-11-27 09:32:07 -06:00
Junegunn Choi
9c798b119c
Remove Python 2.4-incompatible 'with' statement
2013-11-25 17:01:24 +09:00
jctanner
2855684538
Merge pull request #4780 from mscherer/fix_accelerate_various
...
Fix various warning on accelerate module, found by pyflakes
2013-11-19 12:42:32 -08:00
James Tanner
0ce1123693
Fixes #4851 Set syslog debug mode for accelerate properly
2013-11-18 09:56:12 -05:00
John Jarvis
bf435e2032
Adding more wait conditions to the wait_for module
...
Adds a new option to wait_for to wait for a file to exist on the file
system before continuing, optionally takes a search_regex param to match
text in the file.
2013-11-12 19:04:38 -05:00
Michael Scherer
ea6567dbf7
fix incorrect send_data call. there is no client variable, and send_data
...
is a class method. Pyflakes complained about it.
2013-11-07 08:11:56 +01:00
Michael Scherer
eb77ca80df
remove unused variable (pyflakes)
2013-11-07 08:11:56 +01:00
Michael Scherer
061394fff7
make pyflakes a bit more silent
...
- explictely import json, since this is used ( only to silence pyflakes)
- remove shutil as it is unused
2013-11-07 08:11:56 +01:00
jctanner
8e105bbfba
Merge pull request #4772 from mscherer/fix_4763
...
Try/finally is a python 2.5 construct, so not working on EL5. Fix #4763
2013-11-04 15:36:02 -08:00
James Tanner
c1ed47933b
Fixes #4485 add an ipv6 parameter to accelerate so that the daemon will bind to an ipv6 port instead of ipv4
2013-11-04 17:20:03 -05:00
Michael DeHaan
83d50b1984
Make it slightly more clear what side of the connection is missing keyczar if using accelerate mode.
2013-11-03 12:46:21 -05:00
Michael DeHaan
334dcfa2fa
Update module documentation.
2013-11-02 09:34:44 -04:00
Michael Scherer
1ef9bed752
Try/finally is a python 2.5 construct, so not working
...
on EL5. Fix #4763
2013-11-02 11:24:29 +01:00
Michael DeHaan
7f125567cb
Some simplification of include_vars and renamed include_files to include_vars.
2013-10-31 19:24:35 -04:00
Michael DeHaan
b69ab89eb1
Merge branch 'include_files' of git://github.com/bennojoy/ansible into devel
2013-10-31 19:14:27 -04:00
Michael Scherer
8efe53eebb
Fix accelerate module on centos 5 by removing
...
unsupported string prefix in 2.4
http://docs.python.org/release/2.4/ref/strings.html
This is also ignored on later version of python 2
and is present only for python 3 conversion, as seen
on http://docs.python.org/2/reference/lexical_analysis.html#string-literals
Also see http://stackoverflow.com/questions/6269765/what-does-the-b-character-do-in-front-of-a-string-literal
Fix https://github.com/ansible/ansible/issues/4739
2013-10-31 21:04:34 +01:00
bennojoy
b53e7353f9
include files
2013-10-28 17:40:09 +05:30
Michael DeHaan
21d4400c96
Misc pep8 fixes
2013-10-19 13:03:15 -04:00
Michael DeHaan
e131de4de0
Added "debug: var=variableName" capability.
2013-10-14 21:01:38 -04:00
James Cammarata
d21714a37f
Fix for network byte order issues in accelerate
2013-10-02 15:37:15 -05:00
James Cammarata
912e3a7b0b
Merge branch 'accelerate_improvements' into devel
...
Conflicts:
library/utilities/accelerate
2013-10-01 21:22:17 -05:00
James Cammarata
fa80a17aa3
Make recv_data less greedy so it doesn't eat other packets
2013-10-01 16:50:32 -05:00
James Cammarata
59a5ce23d9
Adding an accelerate_timeout parameter for plays
...
This setting makes the timeout for each play configurable, rather than
hard-coding it at 300 seconds (now the default if left unspecified)
Fixes #4162
2013-10-01 15:26:50 -05:00
James Cammarata
f9c87868ac
Added keepalive packets to accelerate mode
...
Commands will now be started up in a separate task from the main
handler thread, so that it can be monitored for completeness while
sending a keepalive packet back to the controller to avoid a socket
receive timeout.
2013-10-01 15:19:21 -05:00