898cf32100
Merge branch 'stable'
2017-04-02 21:57:33 +03:00
ab6ccc5f1f
(c) year
2017-04-02 21:47:58 +03:00
699a7ce01e
Change cluster_size_2 test ordering to fix failing test
...
Signed-off-by: Gerhard Lazu <gerhard@rabbitmq.com>
2017-03-13 12:13:21 +00:00
eb3f004280
Extend notify and notify_clear to support usernames
2017-01-13 15:48:45 +01:00
b5db85e3e0
federation_link_restart => restart_federation_link
...
while at it, provide a more user friendly error message
2017-01-04 02:27:15 +08:00
b1b07fbfaa
CLI command to restart a link
2017-01-03 15:26:11 +01:00
9139e8e43f
Merge branch 'stable'
2017-01-03 15:11:37 +01:00
c8252f2c53
Internal rename of only-down to only_down to satisfy parser
2016-12-30 16:07:33 +01:00
600b29df43
Add tests for federation status/lookup
2016-12-29 10:06:56 +01:00
46e5cddeaf
Rename option
2016-12-28 13:23:33 +01:00
221de38065
CLI command option to list down links
2016-12-28 11:32:12 +01:00
12ad37f1f3
Include empty fields for csv and table formatters
2016-11-21 11:21:34 +00:00
6fb03876c1
Tests for CLI extension for federation status
2016-11-21 10:06:04 +00:00
9b70d870b8
Test refactor
2016-11-21 10:06:04 +00:00
548f65eb6e
Merge branch 'stable' into rabbitmq-federation-39
2016-10-19 17:12:48 +08:00
174b6134b4
Improve Federation link connection name
...
Fixes #39
2016-10-19 09:38:47 +02:00
a4c514f39a
Test cleanup
2016-10-14 14:44:53 +01:00
b180f575a1
Make Federation connections (links) specify a name
...
Fixes #39
2016-10-13 18:01:36 +02:00
dc14525cf4
exchange_SUITE: Increase timetrap to 5 minutes
...
Hopefully this should fix timeouts on CI.
2016-07-06 14:53:00 +02:00
06d248d412
exchange_SUITE: Remove debugging printf
2016-06-24 14:23:29 +02:00
4e70c06538
Switch testsuite to common_test
...
While here, update Erlang.mk. It fixes the use of the `t=` argument to
`make ct-$suite`.
[#121412011 ]
2016-06-24 12:49:37 +02:00
5b04047100
Update (c) info
2016-01-01 12:59:17 +03:00
3836212a3d
Use $RABBITMQ_NODENAME instead of hard-coding nodename
2015-10-19 17:29:55 +02:00
631b280fe6
(c) year
2015-05-24 05:13:37 +03:00
457ab381ef
Use functions from the broker, remove need to test.
2014-06-11 13:10:02 +01:00
be3cd2864a
Remove some dependencies that were invented. Add a test to make sure the helper exchange type gets unregistered / reregistered correctly.
2014-06-11 12:04:31 +01:00
34b5272870
Add a test designed to blow up and remind us if those records change. The alternative would be to move the id/1 functions to rabbit_amqqueue and rabbit_exchange, but they (especially the queue one) are just that little bit too federation-specific to really make sense there.
2014-06-11 11:44:06 +01:00
d770e82a1b
Test queues too.
2014-06-10 13:13:59 +01:00
5bc642aa63
Flesh out test, fix bug.
2014-06-10 12:20:53 +01:00
43a5360fed
Start to test plugin enable / disable
2014-06-10 10:28:37 +01:00
d7ca888acc
Remove some no-longer-needed stuff
2014-05-14 14:30:28 +01:00
85967f6f2f
Port all the in-broker tests.
2014-05-14 14:06:05 +01:00
2b556b38be
Add a mechanism for federation to run multiple nodes in one distributed system, and thus port the rest of the mutli-node tests. There is still a great deal of detritus to be removed...
2014-05-13 18:10:31 +01:00
6a88d64fac
Further WIP-saving.
2014-05-09 13:53:54 +01:00
4b787a9d20
WIP: port a test to use the multi-node framework.
2014-05-08 16:26:52 +01:00
8b1d0105eb
Update copyright for 2014
2014-03-17 17:25:21 +00:00
cc32ab542b
remove unused code & eliminate warning
2014-03-13 12:50:51 +00:00
61a01a752e
Wait for the "goes" key to go away - i.e. for the old internal exchange to be deleted - so that when we check for "stays" we will get the new copy and not inadvertently the old and new copies together.
2014-02-26 14:45:01 +00:00
53ef85ec4e
Another final deracification.
2014-02-20 10:46:12 +00:00
5a5fc63c92
Further attempt to remove raciness from the federation tests. If my calculations are correct, this might be the last one[?!]
2014-02-19 11:43:27 +00:00
ddab02d359
Try to make this test more robust.
2014-02-11 11:25:13 +00:00
524673b392
Merge bug 23906
2014-02-05 16:32:38 +00:00
0382d04e03
Merge bug25979 (again)
2014-02-05 16:26:13 +00:00
9f73f73fdf
Damn, that's no longer relevant either.
2014-02-05 16:24:29 +00:00
933b571aa2
Fix tests post-merge
2014-02-05 16:07:48 +00:00
4e7d0294ed
Merge bug25979 again
2014-02-05 15:30:32 +00:00
4ad9f2b681
That's no longer relevant.
2014-02-05 15:30:18 +00:00
80bfbe2272
Merge default
2014-02-05 15:15:57 +00:00
9e29aa67ea
Derp
2014-01-30 15:08:05 +00:00
523a938917
Remove all the timer:sleep/1 invocations (that aren't part of a loop to check whether something has happened yet). It turns out that they can all be replaced by waiting to see if bindings have appeared upstream. Tests are now notably faster to run, and should be more reliable.
2014-01-30 14:59:52 +00:00
02f59b44e9
Cycle detection for bindings too, plus a test.
2014-01-30 11:17:24 +00:00
ef6f12fef0
Be more general about reporting failed command executions - ?assertCmd() doesn't tell you anything.
2014-01-29 15:21:44 +00:00
299829365a
Update nomenclature in tests too...
2013-12-09 11:18:03 +00:00
36402c280b
Tests need renaming too.
2013-09-18 12:52:04 +01:00
94e4d3293d
Rename a couple of modules that are exchange-specific.
2013-09-18 12:23:28 +01:00
d374681d9e
Merge in default
2013-09-05 12:25:38 +01:00
f5b3a9000e
Merge default
2013-09-04 14:49:23 +01:00
e4addbbc5e
API change
2013-09-03 12:50:09 +01:00
616c4ce764
If we fail to start a second broker, let's say why
2013-09-03 12:46:11 +01:00
cebd0cc414
Merge in default.
2013-09-02 16:36:20 +01:00
b26877f55d
Uh, post bug 25707 that test makes no sense, and fails frequently.
2013-09-02 16:31:32 +01:00
ec17942ffc
Merge default
2013-08-16 11:16:54 +01:00
b5d5996746
Update to new queue decorator API, and thus re-enable a disabled test.
2013-07-05 17:27:27 +01:00
ce515d3310
s/VMware/GoPivotal/g
2013-07-01 10:49:12 +01:00
b81e3f093d
Fix various bugs in status updates, tests now pass again.
2013-06-28 16:46:23 +01:00
ceadd6d480
Add a (simple) restart test and also test status (although that part fails right now).
2013-06-27 16:53:26 +01:00
0cd78b887b
Oops, unbreak tests
2013-06-12 14:51:47 +01:00
ae5907e342
A bunch more tests, and a (temporary) fix thus discovered.
2013-06-05 14:36:11 +01:00
1d4c99e2d6
Disable this test until we are much further along.
2013-06-05 11:56:24 +01:00
844ff645b6
Allow specifying a different queue name for the upstream, remove use of a fake exchange, start to write tests, remember to sepcify drain.
2013-06-04 15:46:43 +01:00
8deff01da2
Add a test of changing the policy for a federated exchange while keeping it federated.
2013-04-16 14:54:41 +01:00
174b3de97c
Stable to default
2013-03-12 12:46:08 +00:00
65a15bbdfa
Oops
2013-03-12 12:34:20 +00:00
4df3ebd76a
Also test that we don't unescape wrongly.
2013-03-12 10:31:11 +00:00
855d654a8f
Rewrite to use uri_parser:parse/2 and add unit tests.
2013-03-12 10:27:59 +00:00
75ae97434a
Strip out credentials when adding URIs to x-received-from header.
2013-03-11 18:44:57 +00:00
6e9e3f4842
Slightly less cheap-and-cheerful test for multiple URIs.
...
"It's all right... to be mean" (to your connection pids).
2013-02-05 13:35:20 +00:00
25dbc11929
Rather fiddly refactoring to split up #upstream into #upstream{} and #upstream_params{} - the latter being instantiated by the link process, and representing a choice having been made amongst the URIs in the #upstream{}. So we now should fail over randomly to a new upstream from the set.
2013-02-04 17:11:39 +00:00
5bd7b6461e
Update copyright 2013
2013-01-23 11:43:00 +00:00
ab318fd339
Merge default!
2012-10-26 16:34:31 +01:00
19ba439616
Correct regexp patterns for tests
2012-10-24 11:40:24 +01:00
b724334ad3
Merged bug25142 into default
2012-10-23 12:45:55 +01:00
cf8ff217e5
s/key/name/
2012-10-22 15:06:37 +01:00
d5f1cee81d
And add a test for binding propagation.
2012-10-18 16:27:47 +01:00
8b3e62d36f
Greater workingness
2012-10-11 18:07:02 +01:00
39d05de40f
Tidying
2012-10-11 10:27:41 +01:00
f70ac3153a
Changes to policy setting interface
2012-10-09 17:21:20 +01:00
64dfa77abc
Updated policy validation
2012-10-05 18:17:03 +01:00
67d3e7b60b
be a bit more careful to ensure nodes are not already active
2012-10-03 16:43:38 +01:00
13b191c689
remove superfluous operator
2012-09-14 18:55:21 +01:00
add2e8591f
user-id tests should use JSON format
2012-09-14 17:24:08 +01:00
b12697b303
Merged default
2012-09-14 12:13:42 +01:00
4cfe3272f3
Merged bug25072
2012-09-13 17:37:31 +01:00
918c8ffd54
increase binding propagation waiting time, to see whether it reduces ci failure rate
2012-09-13 15:24:14 +01:00
82abc37c26
try tracking down ci failures
2012-09-13 13:32:35 +01:00
9eddd1d13e
modified timeout by mistake
2012-09-13 12:20:02 +01:00
823a6cdb12
Merged default
2012-09-13 12:14:38 +01:00
78de6dadda
update the federation tests to use json
2012-09-11 15:00:22 +01:00
9171406883
Make dyn. more obviously a prefix, and remove the priority here too.
2012-09-07 15:29:49 +01:00
e2bc04709c
Add trust-user-id field to upstreams, and test it.
2012-09-07 13:14:35 +01:00
e51d20fb8e
Update federation tests for regex policies
2012-09-06 16:47:36 +01:00
da7663b62b
Rename "federation connection" to "upstream", underscores to hyphens in everything parameter-related.
2012-07-04 12:57:48 +01:00
b14dac68ef
Make all these set_parameter tests more readable.
2012-06-11 16:13:15 +01:00
4d0cc84b00
Simple test for federating / unfederating exchanges at runtime.
2012-06-11 15:51:04 +01:00
d22cbc616b
Get the tests passing again (TODO: write more tests!). This necessitates a bit of rejigging of interfaces (mostly back closer to where it is on default). Note that the validation test is no longer needed :-)
2012-06-08 15:16:40 +01:00
c99c7da468
Merged bug24695 into default
2012-05-28 17:28:01 +01:00
b4bef56044
Don't bother with this tiemout, eunit imposes a timeout on tests anyway and this just provides another way for things to fail when CI is slow.
2012-05-23 14:49:11 +01:00
c7e1a46999
Globally raise timeouts since CI is currently running so slowly that even unit tests are timing out. Hey, it worked for management.
2012-05-16 11:20:28 +01:00
40191251a7
Update the tests to work again (and test direct connections too).
2012-04-24 14:12:49 +01:00
7f5e20067e
Allow exchanges to point to upstream sets that do not exist, and upstream sets to point to connections that do not exist. And tests to match.
2012-04-11 12:07:04 +01:00
7d297e6004
Test dynamic configuration a bit.
2012-04-10 18:09:56 +01:00
c81e107e2c
That's obsolete.
2012-04-10 16:05:28 +01:00
065c11bce0
Damnit
2012-04-05 15:56:29 +01:00
c03564fc92
That's now a binary
2012-04-05 15:33:00 +01:00
a75f498504
No point in testing statically configured exchanges any more
2012-04-05 15:32:43 +01:00
fbb3ef1fb3
Switch to configuring tests via the new config mechanism.
2012-04-05 15:31:54 +01:00
8debf06129
Merge in default
2012-03-23 13:45:27 +00:00
2d1573d652
Fix most tests.
2012-03-23 13:05:06 +00:00
90239e86ad
Simple test for the legacy case where we connect to an upstream that does not know about the x-federation-upstream exchange.
2012-03-13 16:38:59 +00:00
2e7f278554
Merge in default
2012-02-29 16:26:17 +00:00
0857e29636
Use rabbit_basic:append_table_header. This means we need to lose a test, but that code is now tested byt the DLX tests.
2012-02-27 17:46:27 +00:00
fb9cdefe12
Merged bug24624 into default
2012-02-06 16:13:36 +00:00
dd68ca5fd7
Increase timeout to allow tests to pass
2012-02-03 14:58:09 +00:00
f566c42857
Move rabbit_misc:nodeparts/1 and rabbit_misc:makenode/1 into rabbit_nodes
2012-02-02 14:36:16 +00:00
6d297855d5
Update copyright 2012
2012-02-02 14:07:58 +00:00
4444dfe094
Merge from default
2011-11-17 12:00:57 +00:00
38fb367086
Remove debugging aid added in 9bbaae483028.
2011-10-10 15:30:02 +01:00
7dc45636fc
This is failing sometimes in CI, but I can't replicate it here, so let's get some more information out.
2011-10-07 12:02:34 +01:00
c82a26f0b5
Merge bug24095 into default
2011-08-24 11:36:06 +01:00
74cc517eac
This is quite lame, but it does reduce the likelihood of random failures here. On my machine at least :)
2011-08-19 14:13:21 +01:00
c549b5df38
Update to use new rabbitmqctl wait form. Unfortunately this is fiddly because:
...
* Getting the pid file from $$ in the script isn't going to work if we use
erl -detached since, err, it'll start a new process.
* If we background the process then eunit's ?assertCmd macro seems to just
hang. I tried redirecting stdout and stderr, setsid and nohup. Nothing
worked.
* So we have to use os:cmd rather than ?assertCmd. And that doesn't let you
get the status code out! So we have to work around that too...
2011-07-20 16:55:30 +01:00
f89ab6043e
For each test, check the status is correct after bringing up some exchanges. And fix bugs found as a consequence:
...
* keypos is needed in our new table declaration
* we should delete status records when deleting exchange
* add upstream_exchange to status to uniquify
Yes, tests are good, aren't they?
2011-07-15 17:40:22 +01:00
7a4344617c
merge bug23900 into default
2011-07-11 16:40:38 +01:00
4dacf6d0fd
suffix/2.
2011-07-11 14:29:00 +01:00
38779ccf1f
Run binding recovery test on hare to work round bug 24256
2011-07-11 14:02:58 +01:00
061ee38430
One of these just timed out on CI, I *think* for no good reason.
2011-07-08 11:36:26 +01:00
6bb7029078
Test scratch space management in binding_recovery_test also.
2011-07-07 16:37:47 +01:00
26d56c4d2b
Merge in default
2011-07-07 16:09:30 +01:00
e84ba0756f
Port 1f66209756c5 to default from bug 24249
2011-07-07 16:07:10 +01:00
246391e11d
Test our handling of scratch space.
2011-07-07 15:39:45 +01:00
78f845ecd0
Merge in default
2011-07-07 14:40:20 +01:00
916e3f8bdc
Test binding recovery with serialise_events on, since it's hard for rabbitmq-server to do that.
2011-07-06 16:31:34 +01:00
dfb2c458e8
Reduce the scope #exchange{}s get to explore.
2011-07-06 14:55:28 +01:00
20d3f0c511
Attempt to make tests more reliable.
2011-06-24 17:27:27 +01:00
45484c82b0
Run tests without pwd sensitivity, thus fix "make test" in the public umbrella.
2011-06-22 17:52:04 +01:00
6f58f6355e
We don't need such long waits now.
2011-06-22 12:21:37 +01:00
7e844748cd
Reorganise the 100ms delays we use to get round asynchronicity of bindings going upstream.
2011-06-22 12:02:04 +01:00
9698da002d
cosmetic
2011-06-21 19:04:39 +01:00
a83ae8ba79
cosmetic
2011-06-21 17:46:43 +01:00
6cadc015ba
Rename upstream_set argument to upstream-set.
2011-06-21 11:32:45 +01:00
Michael Klishin
Jean-Sébastien Pedron
Diana Corbacho
Michael Klishin
Arnaud Cogoluègnes
Daniil Fedotov
Jean-Sébastien Pédron
Simon MacMullen
Matthias Radestock
Emile Joubert
Tim Watson
Francesco Mazzoli
Rob Harrop