c04f770118
Merge pull request #11506 from SergeTupchiy/EMQX-10274-do-not-download-empty-trace-log
...
fix(emqx_trace): don't download empty trace log file
2023-08-28 22:17:00 +03:00
0072749143
fix(emqx_trace): don't download empty trace log file
...
Closes: EMQX-10274
2023-08-28 19:25:55 +03:00
1e7cde5712
Merge pull request #11447 from thalesmg/cluster-wipe-20230814
...
feat: cluster purge
2023-08-28 12:34:35 -03:00
5908b69353
chore: hide cluster purge from cli and api for now
2023-08-28 10:23:21 -03:00
f988de4ff4
feat(purge): clear delayed messages
2023-08-28 10:23:21 -03:00
9bded07834
feat: cluster purge
...
Fixes https://emqx.atlassian.net/browse/EMQX-10763
2023-08-28 10:23:20 -03:00
dec21ffc95
Merge pull request #11362 from thalesmg/ds-handle-sub-20230725
...
feat(ds): open iterators when handling `SUBSCRIBE` packets
2023-08-28 10:21:28 -03:00
2d931a0512
test(evict): do not disable router module in cluster nodes
2023-08-28 14:07:46 +04:00
8d2ebdea7e
fix(router): generalize config option and make effects visible
2023-08-28 14:07:46 +04:00
e85789306b
chore(router): drop test-only `has_routes/1`
...
Seems that it's too much support work only for test purposes, where
`lookup_routes/1` is nearly as usable.
2023-08-28 14:07:46 +04:00
5d51687dbf
feat(router): add unified routing table config option
2023-08-28 14:07:45 +04:00
33e5e1ba57
feat(router): add unified routing table
2023-08-28 14:07:44 +04:00
166375a000
fix(topicidx): make `get_record/2` simpler to use in concurrent env
...
The mechanic of `emqx_topic_index` cannot really guarantee atomicity
of reading records associated with index matches, so instead it's
probably better to make the user aware of that lack of this guarantee.
2023-08-28 13:52:59 +04:00
84e40fb6fe
test(trie): add more involved route add + delete case
...
That shows how current local trie implementation breaks because
of lack of refcounting.
2023-08-28 13:52:59 +04:00
dcb63440bc
refactor(mgmt): avoid dealing with router tab directly
...
Instead, contain all this behind `emqx_router` module interface.
2023-08-28 13:52:59 +04:00
9eccfa0909
refactor(router): isolate cleanup logic in router module
2023-08-28 13:52:59 +04:00
f3467f44b2
chore: bump hstreamdb_erl driver to `0.4.5+v0.16.1`
2023-08-28 17:28:50 +08:00
4488e9e591
chore(ft): stop hook chain when doing actual handling
2023-08-28 11:57:44 +03:00
39a48179ea
chore(emqx_channel): use macros for reply construction
2023-08-28 11:57:44 +03:00
b8cacd2833
chore(ft): add tests for async reply registry
2023-08-28 11:57:44 +03:00
b74ff10705
Merge pull request #11517 from keynslug/ft/topic-index-v3
...
feat(topicidx): iterate on trie search implementation
2023-08-28 12:24:16 +04:00
d7f6d02270
Merge pull request #11523 from zhongwencool/emqx-conf-suggest-message
...
fix: improve the suggest msg for update conf failed
2023-08-28 14:44:12 +08:00
fe37842b11
chore: update 11523 changelog
2023-08-28 11:27:32 +08:00
053b7fb94a
test: add more test for conf_cli
2023-08-28 11:14:36 +08:00
d8be248a3d
fix: improve the suggest msg for update conf failed
2023-08-28 08:53:23 +08:00
b0d4a22aa8
chore(ft): refactor async reply mechanism
2023-08-27 21:54:23 +03:00
b874926de7
Merge pull request #11514 from zhongwencool/stop-load-application-before-reboot
...
fix: stop otel deps appication before reboot
2023-08-27 20:46:31 +08:00
2bb415f897
Merge pull request #11526 from paulozulato/fix-validator-return-msg
...
fix: flatten error message on resource validator
2023-08-25 17:27:31 -03:00
33a0048155
refactor: move logic to `ensure_iterator`
2023-08-25 15:24:53 -03:00
84e68f8ed8
Merge pull request #11508 from paulozulato/fix-azure-msg-error-handling
...
fix(kafka): fix result handling when sending message with invalid header
2023-08-25 14:57:09 -03:00
42877e282d
fix: flatten error message on resource validator
...
Fixes https://emqx.atlassian.net/browse/EMQX-10864
2023-08-25 13:53:52 -03:00
f15f59650d
test: rm obselete workaround code
2023-08-25 13:49:33 -03:00
cb1e105e19
Merge pull request #11516 from paulozulato/fix-kafka-aeh-client-unique-id
...
Fix Kafka/AEH ClientId uniqueness
2023-08-25 10:14:44 -03:00
b59ec09266
Merge pull request #11520 from zhongwencool/packet-connack-sent-count-error
2023-08-25 21:00:06 +08:00
d5cff533e3
chore(topicidx): drop TODO comment
2023-08-25 16:44:03 +04:00
ca59a87d47
chore(topicidx): refine example of wildcard compare
2023-08-25 16:44:03 +04:00
558402a68e
chore(topic_index): add topic validation
2023-08-25 16:44:03 +04:00
33ed53bb6a
refactor(topicidx): simplify `compare/3` further
...
It's possible to emit seek instruction just once, on the way back
out of the `compare/3` stack.
2023-08-25 16:34:45 +04:00
fc37d235c7
refactor(topicidx): simplify seek instructions
...
Which also avoids comparing filter tail repeatedly when evaluating
it.
2023-08-25 15:24:06 +04:00
00b2712f29
docs: changelog entry for improved too long schema name fix
2023-08-25 11:09:44 +02:00
c65db82b07
fix: bad error message when rule engine schema name is too long
...
Fixes:
https://emqx.atlassian.net/browse/EMQX-10778
2023-08-25 11:02:38 +02:00
c8af5a702c
test: remove connection_closed from test
2023-08-25 15:53:35 +08:00
fc1738188e
fix: packets_connack_sent is not incremented if the ack_flag field in the CONNACK packet is non-zero
2023-08-25 15:18:34 +08:00
ba9cb0a1ed
Merge pull request #11512 from lafirest/fix/ldap_def_port
...
fix(ldap): expose request_timeout to prevent infinite hang
2023-08-25 10:22:44 +08:00
8ab6eb5638
test: start emqx_opentelemetry in emqx_machine SUITE
2023-08-25 10:09:00 +08:00
cf45e80c71
feat(topicidx): iterate on trie search implementation
...
This improves matching performance and decreases GC pressure on
synthetic workloads.
2023-08-25 01:47:11 +04:00
535c7f8b43
fix(kafka): avoid ClientId collision between Kafka and Azure bridges
...
Fixes https://emqx.atlassian.net/browse/EMQX-10860
2023-08-24 18:46:22 -03:00
d9a5a9ea2b
Merge pull request #11513 from thalesmg/kafka-fix-ts-template-r52-20230824
...
fix: use correct timestamp template for kafka and remove such fields from azure event hub producer
2023-08-24 16:26:52 -03:00
4ac0972979
Merge pull request #11501 from thalesmg/gcp-consu-err401-r52-20230823
...
fix(gcp_consumer): handle 401 errors
2023-08-24 15:42:26 -03:00
016ae0524f
fix(aeh_producer): remove timestamp template field
...
Fixes https://emqx.atlassian.net/browse/EMQX-10847
Checking the whole Kafka message from AEH, it seems like the timestamp type is append,
which means that it’s the broker who controls the timestamp, and the timestamp defined by
the producer is ignored.
Ref: https://kafka.apache.org/documentation/#brokerconfigs_log.message.timestamp.type
Example message consumed from AEH:
```
%{
"headers" => %{},
"key" => "",
"offset" => 4,
"topic" => "test0",
"ts" => 1692879703006,
"ts_type" => "append",
"value" => "{\"username\":\"undefined\",\"topic\":\"t/aeh/produ\",\"timestamp\":1692879692189,\"qos\":0,\"publish_received_at\":1692879692189,\"pub_props\":{\"User-Property\":{}},\"peerhost\":\"undefined\",\"payload\":\"aaaa\",\"node\":\"emqx@127.0.0.1\",\"metadata\":{\"rule_id\":\"rule_aehp\"},\"id\":\"000603AA44B34E08F4AF000006E30003\",\"flags\":{},\"event\":\"message.publish\",\"clientid\":\"undefined\"}"
}
```
Note the ts_type above is append.
Example message from a Kafka broker whose ts type is create:
```
%{
"headers" => %{},
"key" => "",
"offset" => 4,
"topic" => "test-topic-three-partitions",
"ts" => 1692881883668,
"ts_type" => "create",
"value" => "{\"username\":\"undefined\",\"topic\":\"t/kafka/produ\",\"timestamp\":1692881883668,\"qos\":0,\"publish_received_at\":1692881883668,\"pub_props\":{\"User-Property\":{}},\"peerhost\":\"undefined\",\"payload\":\"aaaaaa\",\"node\":\"emqx@127.0.0.1\",\"id\":\"000603AAC7529FEEF4AC000007050000\",\"flags\":{},\"event\":\"message.publish\",\"clientid\":\"undefined\"}"
}
```
Unfortunately, I couldn’t find anywhere in AEH where that configuration could be changed.
2023-08-24 14:20:42 -03:00
SergeTupchiy