Commit Graph

1026 Commits

Author SHA1 Message Date
Sean Chittenden faa2aa4436 Guard against a node name containing pure whitespace 2016-03-31 15:02:58 -07:00
Sean Chittenden c03d025903 Node names are not allowed to be empty 2016-03-31 14:47:55 -07:00
Sean Chittenden 85c1da796d Change tests to accept an "at least this many" answers semantic
The size of answers vary based on the target platform.  Accomodate this variance.
2016-03-30 11:48:32 -07:00
Sean Chittenden 12f1a71f00 Fix tests
I'm not sure how I thought this passed last night, it was obviously wrong.
2016-03-30 10:16:05 -07:00
Sean Chittenden 1bcf8d9b33 Editorialize the comment re: UDPAnswerLimit
Prodded by: slackpad & fusiondog
2016-03-30 01:15:40 -07:00
Sean Chittenden 9d7a6d895a Use table-driven test for response limits
Much more exhaustive testing and shows where the limits are of the 512B limitation (quering by ID is less space efficient than querying by just a prepared query or service).
2016-03-30 01:05:49 -07:00
Sean Chittenden 49f8b147b8 Remove duplicate paste-o
Pointed out by: slackpad
2016-03-29 23:31:31 -07:00
Sean Chittenden 8f7bed30a6 Test services and PQT limits 2016-03-29 23:08:58 -07:00
Sean Chittenden 6392dfe0e3 Improve the use of constants across DNS tests 2016-03-29 23:08:40 -07:00
Sean Chittenden b2e0baec4e Use a non-default config value 2016-03-29 23:06:31 -07:00
Sean Chittenden fbf1e5c939 Test missing config inputs
Alpha sort where appropriate to make it easier to keep in sync in the future.
2016-03-29 23:06:17 -07:00
Sean Chittenden c3dcd9acfb Correctly detect that an answer has been truncated 2016-03-29 19:52:31 -07:00
Sean Chittenden 0ad4fab7d4 Remove dead and incorrect code
This functionality is handled further upstream in `trimUDPAnswers()`
2016-03-29 19:38:26 -07:00
Sean Chittenden d1166307aa Trim UDP responses per configuration 2016-03-29 19:27:21 -07:00
Sean Chittenden a29f6e8c8f Allow adjusting the number of DNS records in a response...
Based on work done by @fusiondog in #1583, extend the concept to use an integer instead of a boolean.

Fixes: #1583 && #1481
2016-03-29 19:23:56 -07:00
Patrick Feliciano e533f9bce1 Adding singleton option to DNS for getaddrinfo bug. 2016-03-29 19:23:07 -07:00
Sean Chittenden 74c8da079e Fix whitespace alignment in a comment 2016-03-23 16:00:39 -07:00
James Phillips 2c61c1d333 Merge pull request #1861 from hashicorp/b-flaky-test
Widens coordinate update sleeps in unit tests.
2016-03-21 18:24:05 -07:00
James Phillips 4629871f57 Widens coordinate update sleeps in unit tests. 2016-03-21 18:23:11 -07:00
Hrishikesh Barua bcca7efe83 Added help text for -dev option #1804 2016-03-21 17:27:57 +05:30
Wim 508bc796a8 Allow [::] as a bind address (binds to first public IPv6 address) 2016-03-18 23:59:44 +01:00
Miquel Sabaté Solà 314a0913e2 command: deprecated the -dc flag in the agent CLI
The `-dc` flag from the agent CLI command has been deprecated in favor of
`-datacenter`. This is done this way because:

- Other CLI commands used `-datacenter`. See: event, exec and watch.
- The agent configuration file uses `datacenter`.

Signed-off-by: Miquel Sabaté Solà <msabate@suse.com>
2016-03-18 15:44:35 +01:00
James Phillips dda84c0f58 Merge pull request #1819 from hashicorp/f-check-accept
Adds default set of accept headers for HTTP checks.
2016-03-09 21:12:25 -08:00
James Phillips 30dba02fa3 Adds default set of accept headers for HTTP checks. 2016-03-09 21:11:20 -08:00
James Phillips b132e8f21b Tweaks algorithm so it's safe with an empty list and adds a PQ test. 2016-03-08 23:09:09 -08:00
Igor Dubinskiy f4edb28ef3 Make sure UDP DNS responses aren't larger than allowed 2016-03-07 16:41:17 -08:00
James Phillips cb1bf35ce9 Updates to latest compiled ui assets. 2016-03-07 13:49:08 -08:00
James Phillips 275c84a0cc Renames "debug" endpoint and structures to "explain". 2016-03-07 10:45:39 -08:00
James Phillips 8493640b09 Adds a prepared query debug endpoint. 2016-03-07 10:45:39 -08:00
James Phillips bfe6f073fe Tweaks formatting of inline output messages. 2016-03-02 19:47:00 -08:00
James Phillips d709ef537b Retains the last output when a TTL check times out. 2016-03-02 17:58:01 -08:00
James Phillips bd4f2ee6b7 Adds a new PUT-based TTL check update endpoint. 2016-03-02 17:54:01 -08:00
Stefan Engstrom b20278cadc add accept header */* for agent check 2016-02-19 10:31:00 -06:00
James Phillips a29c1ee04e Merge pull request #1703 from alistanis/fix-issue-#1661
fixes issue #1661 and adds supporting test
2016-02-16 20:13:36 -08:00
James Phillips cbdff8296f Fixes redirect from / to /ui when internal UI is enabled. 2016-02-12 16:11:32 -08:00
Chris Cooper 1327d9eff6 add missing test 2016-02-09 10:49:41 -05:00
Chris Cooper a2533d5d76 fixes issue #1661 and adds supporting test 2016-02-09 10:35:39 -05:00
James Phillips 968bd6321c Adds a sanity check to the local node info compare. 2016-02-07 15:07:23 -08:00
James Phillips 9fabd05157 Merge pull request #1698 from hashicorp/pr-1547-slackpad
Implements WAN address translation.
2016-02-07 14:26:04 -08:00
James Phillips 6797b3ee85 Adds a test to make sure we get the local address within the DC. 2016-02-07 14:16:15 -08:00
James Phillips 3be9f16030 Increases wait to account for random stagger. 2016-02-07 13:56:45 -08:00
James Phillips a61a7403a7 Factors address translation into a single function. 2016-02-07 13:39:37 -08:00
James Phillips f163522f0c Moves tagged wan address to be managed by anti-entropy, not serf. 2016-02-07 13:12:42 -08:00
James Phillips 81e416e70b Adds a test to make sure the local sync doesn't clobber tagged addresses. 2016-02-07 11:26:19 -08:00
James Phillips 79bd1fd4bb Sets up config for more address tags down the road, renames struct members. 2016-02-07 10:37:34 -08:00
Evan Gilman 71cf39b5f8 Use a map for additional node addresses 2016-02-06 23:01:45 -08:00
Evan Gilman a8fd9542dc Add tests for consul WAN translation 2016-02-06 23:01:45 -08:00
Evan Gilman 28ce1b079f Explicitly set advertise address in agent tests
Otherwise, the tests will fail when run on a machine with multiple
private addresses
2016-02-06 23:01:45 -08:00
Evan Gilman d18dc81749 Consider WAN addresses during Node and SRV record lookup 2016-02-06 23:01:45 -08:00
Evan Gilman 0b741350b9 Implement `translate_wan_addrs` config option
This knob tells consul whether it should prefer the WAN address (if set)
when making service lookups in remote datacenters. This enables
reachability for remote services which are behind a NAT.
2016-02-06 23:01:45 -08:00
Evan Gilman d123b4dece Rectify value of `AdvertiseAddrWan` when set elsewhere
`AdvertiseAddrs` has been introduced as a configuration option, which
duplicates a few other options, namely `AdvertiseAddrWan`. We need to
use this value elsewhere, so rather than doing a precedence check every
time we need to access it, rectify the value of `AdvertiseAddrWan` to
match
2016-02-06 23:01:45 -08:00
James Phillips 1a828e3927 Store WanAddress during Service/Check sync 2016-02-06 23:01:45 -08:00
James Phillips 5aedfbac34 Merge pull request #1690 from alistanis/use-http-package-statuses
refactors http server error codes
2016-02-06 22:38:48 -08:00
James Phillips a599e88507 Merge pull request #1284 from nbrownus/telemetry
Option to disable hostnames from telemetry
2016-02-06 22:00:14 -08:00
James Phillips 7273fb0470 Merge pull request #1654 from ryanslade/invalid-check-message
Updated invalid check error message.
2016-02-05 17:21:29 -08:00
Chris Cooper de4129fced Merge branch 'master' of https://github.com/alistanis/consul into use-http-package-statuses 2016-02-05 17:30:43 -05:00
Chris Cooper 0ce79f8198 add comment codes 2016-02-05 17:06:42 -05:00
Chris Cooper 4890220929 refactors http server error codes to use the http statuses defined in the http package instead of literals 2016-02-05 16:51:31 -05:00
Sean Chittenden 5ac7b893cc Use panic instead of returning a sentinel UUID values in unit tests 2016-02-01 16:42:04 -08:00
Sean Chittenden c470553b6b Factor out duplicate functions into a lib package
Consolidate code duplication and tests into a single lib package.  Most of these functions were from various **/util.go functions that couldn't be imported due to cyclic imports.  The consul/lib package is intended to be a terminal node in an import DAG and a place to stash various consul-only helper functions.  Pulled in hashicorp/go-uuid instead of consolidating UUID access.
2016-01-29 16:57:45 -08:00
Nate Brown 541a98c9c7 Option to disable hostnames from telemetry 2016-01-29 13:44:48 -08:00
Ryan Slade 37c7a0e006 Updated invalid check error message.
Added some of the newer reasons this error could have occured.
2016-01-28 11:02:00 +02:00
James Phillips 0fcdd37107 Fixes the static asset generator for new pkg path, updates assets. 2016-01-15 10:21:42 -08:00
Seth Vargo dc9131a732 Use HTTPS + www. where appropriate 2016-01-13 17:44:01 -05:00
James Phillips b21ef7330f Merge pull request #1592 from hashicorp/b-reap
Adds support for the reap lock.
2016-01-13 09:37:54 -08:00
James Phillips d22fe5ba62 Adds support for the reap lock. 2016-01-12 21:10:25 -08:00
Matt McCoy f8a6b6db83 Add consul agent help for advertise-wan 2016-01-11 14:27:54 -05:00
James Phillips df361afed5 Merge pull request #1550 from hashicorp/f-devmode
Add a `-dev` mode for easy prototyping
2016-01-05 10:01:41 -08:00
Philippe M. Chiasson e308bc3733 remove address splitting 2016-01-05 11:49:40 -05:00
Philippe M. Chiasson 6f316138d3 Move logged client IP to the end of the log message 2015-12-30 15:39:34 -05:00
Philippe M. Chiasson 8e938d4d38 Merge branch 'master' of http://github.com/hashicorp/consul into issue/1447/http-access-logs 2015-12-30 15:38:30 -05:00
Ryan Uber f37c0facbe agent: enable ui for dev mode 2015-12-26 20:39:42 -05:00
Ryan Uber ce4cf8a542 consul: dev mode works 2015-12-26 20:19:36 -05:00
Ryan Uber b04510f934 agent: update assetfs output 2015-12-24 22:21:34 -05:00
Ryan Uber 52ad989405 agent: static UI is configurable 2015-12-24 22:21:34 -05:00
Ryan Uber 86f76205aa agent: initial web assets 2015-12-24 22:21:34 -05:00
Ryan Uber a262115679 agent: compile web assets into consul binary 2015-12-24 22:21:34 -05:00
Philippe M. Chiasson b8f8861d97 Merge branch 'master' of http://github.com/hashicorp/consul into issue/1447/http-access-logs 2015-12-22 14:55:18 -05:00
James Phillips 4dd7a5fbaa Adds a Bool helper function. 2015-12-22 10:43:32 -08:00
James Phillips d71036c4f0 Moves logger down where it's used for reaping. 2015-12-22 10:29:55 -08:00
James Phillips e1d456a079 Changes sense of option to "reap" and uses nil for "not set". 2015-12-22 10:28:10 -08:00
James Phillips 1e4aa28c9d Adds child process reaping when Consul is running as PID 1. 2015-12-21 21:47:35 -08:00
James Phillips 874d491ade Converts the DNS metric to a gauge which gives us a count and a time. 2015-12-21 18:25:09 -08:00
James Phillips 9e0ac674c3 Adds telemetry on number of DNS queries served, per-agent. 2015-12-21 17:01:28 -08:00
WangFenjin dcecbf4653 Fix typo 2015-12-08 17:45:01 +08:00
James Phillips b87aeec4e1 Removes the GOMAXPROCS warnings which are obsolete for Go 1.5+. 2015-11-25 17:59:16 -08:00
Philippe M. Chiasson 396df9a0c1 Add client ip to http access logs
Fixes #1447
2015-11-24 15:06:33 -05:00
James Phillips 23536f2b97 Adds Docker checks support to client API.
Also changed `DockerContainerId` to `DockerContainerID`, and updated the agent
API docs to reflect their support for Docker checks.
2015-11-18 07:40:02 -08:00
James Phillips 573f48ed3d Breaks up huge HTTP endpoint functions. 2015-11-17 09:16:04 -08:00
James Phillips a364a02da6 Avoids taking the length again when parsing DNS queries. 2015-11-17 08:40:47 -08:00
James Phillips a1e02996e5 Moves conversion of nil slices up to HTTP layer for prepared queries. 2015-11-15 17:06:00 -08:00
James Phillips 267c352c6d Adds a test to make sure a stale retry terminates. 2015-11-15 17:06:00 -08:00
James Phillips e6ccf5b9ed Returns a 404 from a get or execute of a nonexistent query. 2015-11-15 17:06:00 -08:00
James Phillips c9ef552385 Plumbs the service name back and uses agent-specific TTL settings as a fallback. 2015-11-15 17:06:00 -08:00
James Phillips cad6938474 Adds unit tests for prepared queries and DNS, using existing tests for equivalence. 2015-11-15 17:06:00 -08:00
James Phillips f9c91479ef Adds DNS support for prepared queries (needs tests). 2015-11-15 17:06:00 -08:00
James Phillips 38daaea503 Adds a slightly more flexible mock system so we can test DNS. 2015-11-15 17:06:00 -08:00
James Phillips 978fb443a7 Adds a note about obfuscating query name/ID from the logs. 2015-11-15 17:06:00 -08:00
James Phillips 124410b66c Adds an HTTP endpoint for prepared queries. 2015-11-15 17:06:00 -08:00
James Phillips a33e35fcce Moves DNS over to new shuffle and filter functions. 2015-11-15 17:06:00 -08:00
James Phillips 763d5ea8a7 Fixes nil slices from HTTP endpoints.
These would manifest in the HTTP output as Javascript nulls instead of
empty lists, so we had unintentionally changed the interface while
porting to the new state store. We added code to each HTTP endpoint to
convert nil slices to empty ones so they JSON-ify properly, and we added
tests to catch this in the future.
2015-11-14 21:05:37 -08:00
James Phillips a384baf671 Merge pull request #1293 from talwai/master
Add options to send telemetry to DogStatsD
2015-11-13 09:51:50 -08:00
talwai 01ccad69a1 Kill unused import in command.go 2015-11-13 11:14:15 -05:00
James Phillips 67408fe477 Fixes unit test fail due to expected truncated message. 2015-11-12 18:16:44 -08:00
talwai e6e842d9e7 Add DogStatsd configuration 2015-11-11 14:30:36 -05:00
James Phillips 9107521c17 Fixes a go vet finding. 2015-10-29 12:47:20 -07:00
James Phillips eb4bfa3411 Prevents agents from considering Raft information when doing sync checks. 2015-10-28 14:32:00 -07:00
James Phillips c1305a08ea Makes the version upshift code look at the correct version field. 2015-10-27 14:44:34 -07:00
Diptanu Choudhury 3b5eefb713 Merge pull request #1343 from hashicorp/f-docker-check
Docker Support for Consul Health Checks
2015-10-26 21:00:48 -07:00
Diptanu Choudhury fb0cae7320 Forcing the Env variable to empty while testing the default shell logic 2015-10-26 20:35:13 -07:00
Diptanu Choudhury 9070d12b2f Added a test for selecting shell from env 2015-10-26 20:21:50 -07:00
Diptanu Choudhury 8af6ac8f16 Making an explicit check to test whether a check is of type Monitor 2015-10-26 19:52:32 -07:00
Diptanu Choudhury f8a03508ad Fixed the tests 2015-10-26 18:20:02 -07:00
Diptanu Choudhury 15ff470a79 Added a test to check if we are properly truncating docker exec outputs 2015-10-26 18:06:55 -07:00
Diptanu Choudhury 709b518d03 Added a test for exit code 1 with docker exec 2015-10-26 16:58:09 -07:00
Diptanu Choudhury c8c08586f0 Fixed the Fake Docker client to simulate Exec start failures 2015-10-26 16:54:51 -07:00
Diptanu Choudhury fd6d57ccac Not adding the docker check if we couldn't create the client 2015-10-26 16:45:12 -07:00
Diptanu Choudhury 210b968037 Making sure the script is not empty if it's a docker check 2015-10-26 16:27:48 -07:00
Diptanu Choudhury 0a657b2602 Updated the comment for CheckType 2015-10-26 15:28:13 -07:00
Diptanu Choudhury 2276ebc978 Collect and truncate the output from docker exec 2015-10-26 15:19:35 -07:00
Diptanu Choudhury 0de0abcaec Defaulting to Monitor check 2015-10-26 15:02:23 -07:00
Diptanu Choudhury d2bd1b0c3f Extracted the logic of figuring out the shell and fixing the logic to find out if the check is a Docker check 2015-10-26 15:00:34 -07:00
Diptanu Choudhury f445318732 Added some tests for docker check 2015-10-26 12:59:40 -07:00
James Phillips b50b0a8526 Merge pull request #1296 from hashicorp/f-fast-sync
agent: remove an N^2 check. See #1265
2015-10-26 11:37:52 -07:00
Diptanu Choudhury d478fa9470 Marking the state of a service as critical if the Docker Daemon doesn't respond while running checks 2015-10-26 11:16:11 -07:00
Diptanu Choudhury b35836fdad Adding a debug log to indicate the exit code of failed check 2015-10-26 10:35:51 -07:00
Diptanu Choudhury 1b1518512c Registering the Exec with Docker Daemon everytime the check is invoked 2015-10-26 10:23:57 -07:00
Diptanu Choudhury c4b85010fb Implemented Docker health checks 2015-10-26 10:23:57 -07:00
James Phillips 2a51d55ce3 Runs go fmt (sorts new go-cleanhttp imports). 2015-10-23 17:14:35 -07:00
James Phillips b91e2d3a97 Completes rebase of network coordinates to new memdb. 2015-10-23 15:23:01 -07:00
James Phillips aea2194ce3 Makes the default protocol 2 and lets 3 interoperate with 2. 2015-10-23 15:23:01 -07:00
James Phillips baa223ab4a Runs go fmt after latest rebase. 2015-10-23 15:23:01 -07:00
James Phillips 8e67c37398 Adds a test for the `DisableCoordinate` config. 2015-10-23 15:23:01 -07:00
James Phillips 3f11bfaea4 Adds coordinate of agent to self endpoint. 2015-10-23 15:23:01 -07:00
James Phillips a74bdcba49 Moves sorting up into coordinate endpoint HTTP handlers. 2015-10-23 15:23:01 -07:00
James Phillips 0e7d5871c5 Installs a friendly handler for coordinate endpoints when coordinates are disabled. 2015-10-23 15:23:01 -07:00
James Phillips def8c6087d Fixes config merge fn for disabling coordinates and adds it to JSON. 2015-10-23 15:23:01 -07:00
James Phillips a2bcef00a0 Adds endpoints for raw network coordinates. 2015-10-23 15:23:01 -07:00
James Phillips fecd639b02 Changes ?near=self to a safer ?near=_agent, which is also clearer about what it does. 2015-10-23 15:23:01 -07:00
James Phillips ee1cf1e13b Adds distance sorting to health endpoint. Cleans up unit tests. 2015-10-23 15:23:01 -07:00
James Phillips 1724b9a6be Adds a magic "self" node name to distance queries. 2015-10-23 15:23:01 -07:00
James Phillips 59b710894d Adds tests for HTTP interface. Removes a stray mark. 2015-10-23 15:23:01 -07:00
James Phillips b63909cf67 Adds coordinate sorting support to catalog queries for nodes and service nodes. 2015-10-23 15:23:01 -07:00
James Phillips d7bac0e808 Turns down the coordinate sync rate a bit more. 2015-10-23 15:23:01 -07:00
James Phillips 9ba9a708f6 Scales coordinate sends to hit a fixed aggregate rate across the cluster. 2015-10-23 15:23:01 -07:00
James Phillips d8b8a3719f Simplifies the batching function and adds some comments. 2015-10-23 15:23:01 -07:00
James Phillips f71c79c53f Does some small cleanups based on PR feedback.
* Holds coordinate updates in map and gets rid of the update channel.
* Cleans up config variables a bit.
2015-10-23 15:23:01 -07:00
James Phillips 1222772452 Hardens Consul from bad coordinates from other nodes. 2015-10-23 15:23:01 -07:00
James Phillips acb0dce829 Moves batching down into the state store and changes it to fail-fast.
* A batch of updates is done all in a single transaction.
* We no longer need to get an update to kick things, there's a periodic flush.
* If incoming updates overwhelm the configured flush rate they will be dumped with an error.
2015-10-23 15:23:01 -07:00
James Phillips b6c31bdf2f Flips the sense of the coordinate enable option. 2015-10-23 15:23:01 -07:00
James Phillips 9c069c5031 Removes one more WAN leftover. 2015-10-23 15:23:01 -07:00
James Phillips edb9a119e2 Does a clean up pass on the Consul side. 2015-10-23 15:23:01 -07:00
James Phillips ac4185b888 Merges config changes after rebase. 2015-10-23 15:23:01 -07:00
Derek Chiang 139c9240ea Address comments 2015-10-23 15:23:01 -07:00
Derek Chiang 840474f170 Add a test case 2015-10-23 15:23:01 -07:00
Derek Chiang 23c08aeeb4 Use IndexedCoordinate instead 2015-10-23 15:23:01 -07:00
Derek Chiang 530e73212a Some fixes 2015-10-23 15:23:01 -07:00
Derek Chiang 2ef802b8b3 Fix a comment 2015-10-23 15:23:01 -07:00
Derek Chiang b2cff43bb5 Complete logic for sending coordinates 2015-10-23 15:23:01 -07:00
Derek Chiang 66b210afcb Some fixes 2015-10-23 15:23:01 -07:00
Derek Chiang b5bbe2bcfa Adding tests and stuff 2015-10-23 15:23:01 -07:00
Armon Dadgar d035dbd43b Merge pull request #1318 from daveadams/f-http-header-token
Allow specifying Consul token in an HTTP request header
2015-10-22 13:33:47 -07:00
Jeff Mitchell 9267f956a2 Update cleanhttp repo location 2015-10-22 14:14:22 -04:00
Jeff Mitchell 06bb9d5f36 Use cleanhttp to get rid of DefaultTransport 2015-10-22 10:47:50 -04:00
David Adams 5f175add40 Add HTTP request header X-Consul-Token
Add support for an X-Consul-Token HTTP request header to specify the
token with which this request should be fulfilled. The header would have
precedence over the responding Agent's default token, but would have
lower precedence than a token specified in the query string.
2015-10-19 11:26:01 -05:00
James Phillips 263c7e3fd3 Deletes the old state store and all its accoutrements. 2015-10-15 14:59:09 -07:00
James Phillips bcdabe4606 Knocks out the Raft indexes before doing compare. 2015-10-15 14:59:09 -07:00
James Phillips cbcd977a39 Gets new structs changes to compile, adds some corner case handling and extra unit tests. 2015-10-15 14:59:09 -07:00
Ryan Uber b46f878747 Merge pull request #1309 from hashicorp/f-remove-migrate
Removes consul-migrate for 0.6
2015-10-15 14:50:19 -07:00
Jeff Mitchell 9cddc187b5 Don't use http.DefaultClient
Two of the changes are in tests; the one of consequence is in the API.
As explained in #1308 this can cause conflicts with downstream programs.

Fixes #1308.
2015-10-15 17:49:35 -04:00
Ryan Uber aba1b26015 agent: consolidates data dir checker 2015-10-15 14:21:35 -07:00
Ryan Uber 8bc51eb237 agent: test mdb dir protection 2015-10-15 14:15:41 -07:00
Ryan Uber 2a7609d6bc agent: remove migrator, refuse to start if mdb dir found 2015-10-15 14:15:08 -07:00
Armon Dadgar 0b77960349 agent: remove an N^2 check. See #1265 2015-10-12 20:30:11 -07:00
Michael Puncel a94589ad67 Add http method to log output 2015-10-02 18:33:06 -07:00
James Phillips 26eadcd95c Merge pull request #1235 from wuub/master
fix conflict between handleReload and antiEntropy critical sections
2015-09-17 07:28:39 -07:00
Wojciech Bederski 9a1b52171f panic when unbalanced localState.Resume() is detected 2015-09-17 11:32:08 +02:00
Dale Wijnand c5168e1263 Fix a bunch of typos. 2015-09-15 13:22:08 +01:00
James Phillips b25797a808 Merge pull request #1187 from sfncook/enable_tag_drift_03
Enable tag drift 03
2015-09-11 15:35:32 -07:00
Anthony Scalisi 8d733b7fca remove various typos 2015-09-11 12:29:54 -07:00
Wojciech Bederski 4cd1b09ad7 make Pause()/Resume()/isPaused() behave more like a semaphore
see: https://github.com/hashicorp/consul/issues/1173 #1173

Reasoning: somewhere during consul development Pause()/Resume() and
PauseSync()/ResumeSync() were added to protect larger changes to
agent's localState.  A few of the places that it tries to protect are:

- (a *Agent) AddService(...)      # part of the method
- (c *Command) handleReload(...)  # almost the whole method
- (l *localState) antiEntropy(...)# isPaused() prevents syncChanges()

The main problem is, that in the middle of handleReload(...)'s
critical section it indirectly (loadServices()) calls  AddService(...).
AddService() in turn calls Pause() to protect itself against
syncChanges(). At the end of AddService() a defered call to Resume() is
made.

With the current implementation, this releases
isPaused() "lock" in the middle of handleReload() allowing antiEntropy
to kick in while configuration reload is still in progress.
Specifically almost all services and probably all check are unloaded
when syncChanges() is allowed to run.

This in turn can causes massive service/check de-/re-registration,
and since checks are by default registered in the critical state,
a majority of services on a node can be marked as failing.
It's made worse with automation, often calling `consul reload` in close
proximity on many nodes in the cluster.

This change basically turns Pause()/Resume() into P()/V() of
a garden-variety semaphore. Allowing Pause() to be called multiple times,
and releasing isPaused() only after all matching/defered Resumes() are
called as well.

TODO/NOTE: as with many semaphore implementations, it might be reasonable
to panic() if l.paused ever becomes negative.
2015-09-11 18:28:06 +02:00
Wojciech Bederski 24ac26b3c1 failing test showing that nested Pause()/Resume() release too early
see: #1173 / https://github.com/hashicorp/consul/issues/1173
2015-09-11 17:52:57 +02:00
Shawn Cook 99be758411 Rename EnableTagOverride and update formatting 2015-09-11 08:35:29 -07:00
Shawn Cook f448a62826 Remove debug lines 2015-09-11 08:32:59 -07:00
Shawn Cook 2f04917261 Merge remote-tracking branch 'hashicorp/master' into enable_tag_drift_03 2015-09-10 14:55:30 -07:00
Shawn Cook 8a86eee9fb Add test cases TestAgentAntiEntropy_EnableTagDrift 2015-09-10 14:08:16 -07:00
Ryan Uber 08d12e978f Merge pull request #1230 from hashicorp/f-maintfix
Respect tokens in maintenance mode
2015-09-10 12:30:07 -07:00
Ryan Uber 948bd57d6a agent: testing node/service maintenance using tokens 2015-09-10 12:08:08 -07:00
Ryan Uber e129a59316 agent: thread tokens through for maintenance mode 2015-09-10 11:43:59 -07:00
Wim 3d7c3725d8 Allow AAAA queries for nodeLookup 2015-09-08 16:54:36 +02:00
Wim 2336c6a4bd No NXDOMAIN when the answer is empty 2015-09-02 16:12:22 +02:00
Ryan Breen a013095f62 Merge pull request #1167 from railsguru/master
Add -http-port option to change the HTTP API port
2015-09-02 01:15:55 -04:00
Armon Dadgar 655666170a agent: Always enable the UI endpoints 2015-09-01 18:28:32 -07:00
Wim e97973c1e1 Limit the DNS responses after getting the NodeRecords 2015-09-01 23:23:05 +02:00
Ryan Breen 56d2fa4c17 Merge pull request #1195 from 42wim/fix-rfc2308-part2
Return SOA/NXDOMAIN when the answer is empty
2015-09-01 17:08:31 -04:00
Wim b806aceef4 Return SOA/not found when the answer is empty 2015-09-01 22:28:12 +02:00
Ryan Uber d6b71de3f4 agent: reload SCADA client if endpoint changes 2015-08-27 13:29:07 -07:00
Ryan Uber 5bd7a5f239 command: atlas endpoint can be passed 2015-08-27 11:11:05 -07:00
Ryan Uber cda2bf6975 agent: atlas_endpoint is configurable 2015-08-27 11:08:01 -07:00
Ryan Uber 00d78963bf agent: log a message when making a new scada connection 2015-08-25 21:03:16 -07:00