- Improve Proxy protocol error logging (warn instead of crash process).
- Add retain command to vmq-admin usage output
- Bridge Plugin: Continue to publish during netsplit (that is, during cluster not_ready)
- Bridge Plugin: Make internal publish use the configured per-topic QoS
- Upgrade package
bcrypt
to fix compilation in OSX (#1500). - Fix issue with loading status dashboard from behind a proxy with a basepath set
- Fix bug where the server would not parse multiple WebSocket protocols correctly.
- New feature: Allow configuration of allowed ECC named curves in configuration file. (thanks to @varnerac)
- New feature: Add
on_topic_unsubscribe
hook (PR #1539, Issue #1326). (thanks to @khodzha) - Update to jquery 3.5.1 and bootstrap 4.1.2 in status page.
- Add case clause for proxy protocol 'local command', causing unnecessary error logging in loadbalancer health checks.
Patch Release to:
- Fix the new bridge metrics labels that prevented the Prometheus metrics to be exported correctly (blocking Prometheus export entirely).
- Improve buffering in in-memory queue of outgoing bridges.
- Add a Name to bridges, so that a bridge can be identified by {Name, Host, Port}, instead of {Host, Port} only. This allows multiple bridges to the same remote endpoint.
- Add bridge Name and MQTT client mailbox size to
vmq-admin bridge show
command. - Add per QoS/per bridge in and out counters to pull resp. push bridges.
- Log bridge connection setups and subscribes (info level).
- Support Power Linux LE (ppc64le) platform.
- Add
on_session_expired/1
hook tovmq_webhooks
schema. - Add Subscriber trie/event handler readiness check to handle fast subscribers after a reboot (race condition, #1557).
- Update nodetool escript with latest version.
- Fix internal Watermark update in SWC (#1556).
- Handle LevelDB truncated record corruption more automatically in LevelDB SWC and Messagestore backends.
- Catch ETS table call in
vmq_cluster:netsplit_statistics/0
(#1537). - Add compatibility with Erlang/OTP 23. This release requires Erlang/OTP 21.3 or later.
- Upgrade package
bcrypt
to version 1.1.0. - Upgrade package
hackney
to version 1.16.0 (dependencies ofhackney
were updated as well). - Fix to allow equal signs on parameter values on
vmq-admin
commands (#740).
URL parameters on webhooks
- Add new
on_session_expired/1
hook tovmq_diversity
andvmq_webhooks
. - Add datetime prefix to the tracer output. The datetimes are expressed in UTC with iso-8601 format (eg '2020-04-29T21:19:39Z'). (#782)
- Add start command to Bridge CLI.
- Fix options passed to PublishFun for plugins (#1516)
- Add configurable connect time (mqtt_connect_timeout) between establishing the connection and sending CONNECT (#735, #824).
- Fix tracing error when using payload modifiers (#1517).
- Make
vmq-admin bridge show
command more robust against bridge client state machine timeouts (#1515). - Fix vmq health http format error (#1529).
- Add new CLI output formatter (#1502).
- Remove minor functions for OTP 18. OTP 18 is not supported anymore (#1523).
- Update VerneMQ Schema info on SWC.
- Add Inet6 support to the vmq_diversity MySQL plugin (#1461).
- Improve error message when plugin chains have exhausted (#1465).
- Upgrade clique dependency to fix bug with empty KV params.
- Support multilevel bridge prefixes.
- Make SSL cert depth configurable for bridges.
- Fix directory paths for metadata backend for RocksDB and Leveled.
- Fix number only client IDs in vmq-admin session CLI (#1436).
- Fix proxy protocol handling bug for WebSockets listeners (#1429).
- Fix compilation for OS X.
- Fix bug in vmq_swc where a cluster leave didn't properly cleanup the node clock.
- Upgrade the Plumtree metadata backend to include a fix for a crash which could occur during metadata exchange with a remote node if the remote node becomes unavailable.
- Fix bug where vmq_metrics crashes because external metric providers haven't started yet.
- Fix compilation on OSX Mojave and Catalina. Requires running
brew install openssl snappy
.
- Fix bug where websocket MQTT clients that didn't provide a
sec-websocket-protocol
header to cause a crash to be logged. This case is now handled and the connection is terminated gracefully (#1317). - Fix bug which caused clients MQTT bridges using protocol versions (131 and 132) to be unable to connect (#1306).
- Upgrade the
hackney
so VerneMQ is compatible with a TLS change in Erlang/OTP 22.1. - Fix bug preventing MQTT 5 publish in/out counts from being shown on the HTTP status page.
- Fix lager issue on Raspberry Pi preventing VerneMQ from starting (#1305).
- Upgrade
epgsql
dependency to version 4.3.0 to get better error messages (#1336). - Add new
max_last_will_delay
valueclient
which means the value from the client is used instead of being overriden by the broker. Thisclient
value is now the default. This setting only applies to MQTT 5.0 clients. - Add hidden option
response_timeout
tovmq_webhooks
endpoints. With this the timevmq_webhooks
waits for a response from the remote endpoint can be configured. Default is 5000 milliseconds. - Add new
on_deliver/6
andon_deliver_m5/7
hooks which adds QoS and retained information to the parameters. The old variants are deprecated and will be removed in the next major version. - Add new
on_session_expired/1
hook which is called when an offline session expires and the state is deleted. - Handle CRL PEM entries with certificates containing empty revocation lists (#1337).
- Fix typo in
vmq-admin listener start
command (#1348). - Optimize subscription performance when many retained messages are stored on the broker and the subscription contains wildcards.
- Fix bug where MQTT 5 publish expiry interval was lost when writing the message to the message store.
- Fix bug where a retried MQTT publish after resuming an offline session used the wrong message id.
- Fix MQTT 5 shared subscription bug where writing the message to the message store resulted in a corrupt message record once a offline session was resumed.
- Refactor built-in message storage to enable multiple storage engines as well as to streamline the implementation of alternative message storage plugins.
- Fix metric description string in vmq_bridge.
- Fix regression in vmq_bridge where the retain bit wasn't handled properly for incoming publishes.
- Enable to define the MQTT protocol version to be used by vmq_bridge (4 or 3).
- Fix vmq_bridge dynamic configuration handling.
- Fix multiple minor issues in vmq_swc.
- Fix HTTP handlers to support parameterized media types.
- Upgrade lager, poolboy, jsx, luerl dependency.
- Fix bug causing idle websocket connections to be closed after 60 seconds (#1292).
- Fix MQTT 5.0 bug causing LWT not being sent when client disconnected with Disconnect with Will Message (0x04) (#1291).
- Ensure MQTT 5.0 subscription identifers are added to messages delivered via shared subscriptions (#1294).
- Fix protocol version check bug which would allow MQTT 5.0 clients to connect even if the MTT 5.0 protocol was not specified for the listener. This happened for clients which connected with an empty client-id.
- Fix bug in
vmq_diversity
andvmq_webhooks
where the plugin hooks would be registered out of order, even though the plugins themselves would be started in the correct order (#1295).
- Update
cuttlefish
to fix parse issue with lines in thevernemq.conf
file consisting solely of white-space (#1208). - Fix bug in the
vmq_webhooks
auth_on_subscribe_m5
andon_subscribe_m5
hooks (#1280). - Fix issue where errors would be logged when the /health endpoint was called (#1281).
- Ensure MQTT 5.0 subscription identifiers are added to retained messages delivered to the client when the subscription is made (#1287).
- Fix bug in
vmq_swc
which could prevent VerneMQ from starting up.
- Ensure mountpoints specified at the protocol level are inherited on the specific listeners.
- Fix missing output from
vernemq version
(#1190). - Ensure errors from the parser variable decode are handled correctly and logged at the debug level instead of causing a crash in the socket process.
- Handle list of
sec-websocket-protocol
correctly (#1149) in the websocket implementation. - Ensure that plumtree application isn't started when metadata plugin SWC is used.
- Fix bug preventing restarting Lua states properly in case of a crashing script.
- Fix warnings due to deprecated lager configuration (#1209).
- Upgrade
lager
to version 3.7.0 (In preparation for OTP 22 support). - Make VerneMQ run on Erlang/OTP 22.
- Add new webhooks CLI register flag
--no_payload=true
which, if enabled, will remove the MQTT payload from the JSON object send via theauth_on_publish
andauth_on_publish_m5
payloads. The flag can also be set in the config file usingvmq_webhooks.hookname.no_payload=on
. - Add metric
client_keepalive_expired
which tracks clients that failed to communicate within the keepalive time. - Expose the
crypto:hash/1
function in LUA, for example to do a 256bit sha3 hash one can callcrypto.hash("sha3_256", data)
. The hashing algorithms supported are exactly the ones supported by thecrypto:hash/1
function and may change depending on the Erlang/OTP version used to build Erlang. See the complete list of supported hashes here: http://erlang.org/doc/man/crypto.html#Digests%20and%20hash. If passed an unknown hashing algorithm an error is raised. - Fix prefix handling in the bridge plugin (vmq_bridge).
- Strengthen parameter validation in the
bcrypt.hashpw/2
LUA function invmq_diversity
. - Pass arguments correctly to
vmq-admin
when called viasudo
. - Handle rate metric labels correctly in the example grafana dashboard.
- Upgrade cowboy to version 2.6.3 as well as cowlib and ranch to versions 2.7.3 and 1.7.1 respectively. This update also means the PROXY protocol code was removed and the PROXY support from cowboy is used instead.
- Handle retained flag in the bridge plugin (vmq_bridge).
- Ensure systemd doesn't terminate VerneMQ if it is slow to start.
- Implement bridge protocol handling (protocol version 131) for bridges connecting to VerneMQ such that the retained bit is kept on messages routed to the bridge (Retained as Publish) and messages coming from the bridge are not forwarded to the bridge if it has a matching description (No Local).
- Fix message store startup consistency checking routine which resulted in potentially deleting wrong messages in cases where an inconsistency was detected. This fix also increases message storage performance during startup as well as during normal operation. However, the nature of the bug and the provided solution don't enable a simple downgrade path and introduces a backward incompatibility if a VerneMQ message store containing offline messages has to be downgraded to 1.8.0 or earlier versions.
- Upgrade dependency
cuttlefish
to version 2.2.0. - Improve large queue initialization performance by reducing algorithmic complexity from O(n^2) to O(nlogn) where n is the number of offline messages.
- Add the ability to modify the
username
onauth_on_register
andauth_on_register_m5
hooks. Supports bothvmq_diversity
andvmq_webhooks
. - Upgrade the
vmq_diversity
redis drivereredis
to version 1.2.0. - Fix
vmq_diversity
supervisor restart shutdown issue (#1241).
- Cleanup cluster state information on a node which is being gracefully removed from the cluster so if it is restarted it comes back up as a stand-alone node and won't try to reconnect to the remaining nodes.
- The
allow_multiple_sessions
option is deprecated and will be removed in VerneMQ 2.0. - Add new
coordinated_registrations
config option which allows for faster uncoordinated registrations (replaces side-effect ofallow_multiple_sessions
). - Upgraded dependency
swc
. Improves memory and cpu usage by fixing node clock update bug. - Set queue expiration (
persistent_client_expiration
) timer on queues after a broker restart (#1071). - Fix status page to show all cluster nodes (#1066).
- Fix Lua
json.decode()
when decoding empty Json objects (#1080). - Fix
vmq_diversity
cache invalidation timing issue when a client reconnects right after a disconnect (#996). - Fix retry of MQTT publish frame in
vmq_bridge
(#1084). - Fix outgoing qos2 message id bug (#1045).
- VerneMQ now requires Erlang/OTP 21.2 or newer to build.
- Use
atomics
andpersistent_term
to handle metrics instead of using themzmetrics
library which has been removed. - Change
vmq_reg:publish/4
to return{ok, {NumLocalMatches, NumRemoteMatchs}}
instead ofok
. This introduces a breaking change to the unofficial publish API provided by thevmq_reg:direct_publish_exports/1
. - Add the
router_matches_local
androuter_matches_remote
counter metrics, counting the total number of matched local and remote subscriptions. - Add a routing score to every cluster node listed on the VerneMQ status page.
The routing score consists of two percentages similar to
75 / 25
indicating that 75% of the received MQTT publish frames were routed to local subscribers and 25% were forwarded to subscribers on different cluster nodes. The routing score can be used to detect imbalances in a VerneMQ cluster. - The MQTT 5.0 plugin hook definitions have been updated as follows:
- The
Payload
andProperties
arguments to theon_deliver_m5
hook have been swapped to be consistent with all the other_m5
hooks. - property modifiers are now using the same names as the properties in the
incoming properties, for example the
user_property
is nowp_user_property
. - Property modifiers have been moved into a
properties
modifier map. So for example instead of returning{ok, #{p_user_property => ...}}
one now returns{ok, #{property => #{p_user_property => ...}}}
. This change makes it possible to drop properties from a plugin hook. - In the
vmq_webhooks
plugin property modifiers have been moved into a separate JSON object with the keyproperties
inside the modifiers JSON object. Thevmq_webhooks
. - The
{ok, binary()}
return clause has been removed from theon_deliver_m5
hook, use the modifier map ({ok, Modifiers}
) instead. - The
{ok, binary()}
return clause has been removed from theauth_on_publish_m5
hook, use the modifier map ({ok, Modifiers}
) instead. - Note, MQTT 5.0 support in VerneMQ is still in Beta.
- The
- Fix error when tracing clients connecting with a LWT.
- Add file validation to check if files in the
vernemq.conf
exist and are readable. - Fix that disconnects due to client migrations were logged as warnings. They are now treated as the normal termination case.
- Improve MQTT 5.0 support in the
vmq_webhooks
plugin: support more properties and improve the tests. - Instead of using the
node_package
dependency to build binary packages, the external Ruby command line toolfpm
is used. This enables cleaning up the Makefile as well as the upgrade of therebar3
build tool. - Improve MQTT 5.0 support in the
vmq_diversity
plugin: support all MQTT 5.0 hooks in Lua and more properties. - Support TLS encrypted connections to PostgreSQL when using
vmq_diversity
. Fixes #811. - Support TLS encrypted connections to MongoDB when using
vmq_diversity
. - Add CockroachDB support to the
vmq_diversity
plugin. - Make it possible to configure how many concurrent bcrypt operations are
possible via the
vmq_bcrypt.pool_size
config. - Fix incorrect format string in shared subscription debug log.
- Upgrade
hackney
to version 1.15.1 (dependencies ofhackney
were updated as well). - Fix bug which could happen while repairing dead queues in case the sync request fails.
- It is now possible to configure TCP send and receive buffer and user-level buffer sizes directly on the MQTT listeners or protocol levels (currently only TCP and TLS listeners).
- Add the
socket_close_timeout
metric. The metric counts the number of times VerneMQ hasn't received the MQTT CONNECT frame on time and therefore forcefully closed the socket. - Fix the potential case of late arriving
close_timeout
messages triggered by thevmq_mqtt_pre_init
. - Change log level from
debug
towarning
when VerneMQ forcefully terminates a MQTT session FSM because of an unexpected message. - Fix
vmq_diversity
ACL rule hash collision issue (#1164). - Improve performance of shared subscriptions when using many subscribers.
- Add the
vmq_pulse
plugin. This plugin periodically sends cluster information to a pulse backend via HTTP(S). The plugin is NOT enabled by default, and once enabled it has to be setup withvmq-admin pulse setup
. This enables a support crew to analyze the cluster performance without having direct access to the VerneMQ nodes. Such a 'Pulse' contains the following information:- Cluster status, similar to
vmq-admin cluster show
- Plugins, similar to
vmq-admin plugin show --internal
- Metrics, similar to
vmq-admin metrics show
- Names and versions of loaded OTP applications
- Erlang System Version
- OS Kernel information, output of
uname -a
- CPU, RAM, and disk usage
- Pulse Version
Note: The
vmq_pulse
plugin is in Beta.
- Cluster status, similar to
- Fix the systemd service unit file to set
LimitNOFILE=infinity
. This enables VerneMQ to request as many file descriptors as configured for thevernemq
user.
- Fix
vmq_webhooks
issue where MQTTv5 hooks where not configurable in thevernemq.conf
file. - Support shared subscriptions in
vmq_bridge
. - Fix bug in Prometheus output (#923).
- Fix
max_message_rate
to include MQTTv5 sessions. - Support new
throttle
modifier for theauth_on_publish
andauth_on_publish_m5
hooks which will throttle the publishing client by waiting for a given number of milliseconds before reading new data from the client socket. Note, backpressure is not supported for websocket connections. This feature was kindly sponsored by Arduino SA (https://www.arduino.cc/). - Fix issue with long-running
vmq-admin
commands (#644). - Added a new HTTP module
vmq_health_http
exposing a/health
endpoint that returns 200 when VerneMQ is accepting connections and joined the cluster (for clustered setups) or returns 503 in case any of the two conditions are not met (#889). - Fix issue where a QoS2 message would be republished if a client would resend the PUBLISH packet with the same message id within a non-finished QoS2 flow (#944).
- Make VerneMQ build on FreeBSD (11.2-RELEASE) with
gmake rel
. - Fix issue with blocking socket commands in inter-node communication.
- Improve error messages when parsing invalid CONNECT packets.
- Log IP and port information on authentication failures to make it easy to block the client with tools like Fail2ban (#931).
- Fix issue which would crash the session if a client would resend a (valid) pubrec during a Qos2 flow (#926).
- Fix
vmq_diversity
error message if invalid or unknown modifiers are returned from a lua hook implementation. - Fix issue preventing MQTT 5.0 properties from being modifiable in
auth_on_publish_m5
(#964). - Fix issue which causes a crash if not enough data is available while parsing
the CONNECT frame in
vmq_mqtt_pre_init
(#950, #962). - Fix issue which could cause
vmq-admin session show
to crash when using--limit=X
to limit the number of returned results (#902). - Handle edge case in the websocket implementation which could cause warninigs when the session was terminating.
- Expose
vernemq_dev_api:disconnect_by_subscriber_id/2
in luavmq_api.disconnect_by_subscriber_id/2
, an example looks like:vmq_api.disconnect_by_subscriber_id({mountpoint = "", client_id = "client-id"}, {do_cleanup = true})
. - Improve
vmq_webhooks
errors so the error from the endpoint is shown as the error reason where relevant. - Optimization: optimize away read operation during client registration by reusing already available data.
- Enable plugins to expose metrics via the available metric providers.
- Add histogram metric type.
- Fix incorrect debug log message in QoS 2 flow.
- Expose bridge metric for messages dropped in case the outgoing queue is full.
- Log (debug level) when a LWT is suppressed on session takeover.
- Optimize subscribe operation by refactoring away one read from the metadata store.
- Add protection mechanism for the plumtree metadata store from subscription
floods by dropping and not acknowledging
i_have
messages when the mailbox size grows above a certain threshold. This threshold is configurable via theplumtree.drop_i_have_threshold
hidden option. The default is 1000000. This work was kindly contributed by ADB (https://www.adbglobal.com). - Add new
vmq-admin retain
commands to inspect the retained message store. - Support for different MySQL password hashing methods in
vmq_diversity
, ensuring compatibility with MySQL 8.0.11+. The method is configurable via thevmq_diversity.mysql.password_hash_method
option which allows:password
(default for compatibility),md5
,sha1
orsha256
. - Fix the HTTP
/status.json
endpoint to have a valid JSON output (#786). - Fix bug where internal application plugins where shown as normal plugins.
- Fix crash in bridge when calling
vmq-admin session show
by fixing thevmq_ql
row initializer to handle plugin sessions (the bridge starts a local plugin session). - Add improvements to VerneMQ status page to have nicer UI and be readable on smaller devices.
- Upgraded dependency
sext
to 1.5.0 (required for better OSX compilation). - Do not accept new client connections while the broker is shutting down as this could cause errors to be reported in the log (#1004).
- Fix
vmq_diversity
PostgreSQL reconnect issue (#1008). - Fix
vmq_webhooks
so peer port is not considered when caching theauth_on_register
to avoid cache misses. - Multiple bug fixes and improvements in
vmq_swc
. - Add the
vmq_swc
metadata plugin (using the already existing LevelDB backend) to the default VerneMQ release. To usevmq_swc
instead ofvmq_plumtree
setmetadata_plugin = vmq_swc
invernemq.conf
.vmq_swc
is still in Beta. - Add missing increments of the
mqtt_connack_sent
metric for CONNACK success(0) for MQTT 3.1.1. - Handle edge case with unknown task completion messages in
vmq_reg_sync
after a restart. - Fix bug which could cause a queue cleanup to block indefinitely and cause the
vmq_in_order_delivery_SUITE
tests to fail. - Add a new metric (queue_initialized_from_storage) to better monitor queue initialization process after a node restart.
- Fix edge case where an extra queue process could be started when metadata events arrive late. Now local queue processes are only started when triggered via a new local MQTT session.
- Reimplement dead queue repair mechanism.
- Add feature to reauthorize existing client subscriptions by reapplying current
auth_on_subscribe
andauth_on_subscribe_m5
hooks. This feature is exposed as a developer API invernemq_dev
, via thevmq-admin session reauthorize
CLI command, and as a API forvmq_diversity
Lua scripts. This work was kindly sponsored by AppModule AG (http://www.appmodule.net). - Improve planned cluster leave queue migration speed significantly (#766).
- Start metrics server before setting up queues to ensure queue metric counts are correct when restarting VerneMQ.
- Let Travis CI build the VerneMQ release packages.
- Add new metric
system_process_count
which is a gauge representing the current number of Erlang processes. - Add
queue_started_at
andsession_started_at
information to thevmq-admin session show
command. Times are POSIX time in milliseconds and local to the node where the session or queue was started.
-
Fix issue when calling a function in a Lua script that requires more time to complete than the default
gen_server
timeout (#589). -
Silently drop messages published by the client that use a routing key starting with '$'.
-
Full MQTTv5 support
With this release VerneMQ officially supports MQTT protocol version 5.0. The MQTTv5 support is currently marked as in BETA and we may still have to change some things.
VerneMQ has full support for the complete MQTTv5 spec, but to list a few of the new features:
-
Support for enhanced (re)authentication
-
User properties
-
Message Expiration
-
Last Will and Testament delay
-
Shared subscriptions
-
Retained messages
-
Request/Response flows
-
Topic aliases
VerneMQ supports topic aliases from both the client to the broker and from the broker to the client.
When a client connects to the broker, the broker will inform the client of the maximum allowed topic alias using the topic alias max property on the CONNACK packet (if it has been set to a non-zero value). The topic alias maximum property can be configured through the
topic_alias_max
configuration variable or overriden in a plugin in theauth_on_register
hook. The broker will then handle topic aliases from the client as per the MQTTv5 spec. -
Flow control
-
Subscription flags Retain as Published, No Local and Retain Handling.
-
Subscriber ids
By default MQTTv5 is disabled, but can be enabled on a listener basis, for example
listener.tcp.allowed_protocol_versions=3,4,5
would enable MQTT version 3.1, 3.1.1 and 5.0 on the TCP listener.MQTTv5 support has been added to the
vmq_passwd
,vmq_acl
,vmq_webhooks
plugins as well as the parts of thevmq_diversity
plugin to support the MySQL, PostgreSQL, Redis and MongoDB authentication and authorization mechanisms. The Lua scripting language provided by thevmq_diversity
plugin does not yet expose all MQTT 5.0 plugin hooks.The
vmq_bridge
plugin currently has no support for MQTTv5.!! Note !! that all MQTTv5 related features and plugins are in BETA and may still change if needed.
-
-
The metrics:
mqtt_connack_not_authorized_sent mqtt_connack_bad_credentials_sent mqtt_connack_server_unavailable_sent mqtt_connack_identifier_rejected_sent mqtt_connack_unacceptable_protocol_sent mqtt_connack_accepted_sent
Have been merged into a single metric
mqtt_connack_sent
and the various MQTT 3.1.1 return codes have been mapped into labels. So for examplemqtt_connack_sent
with the labelreturn_code=success
replacesmqtt_connack_accepted_sent
. -
Added on_message_drop hook that is called for every message dropped due to exceeding the MQTTv5 max_packet_size property, hitting the message expiry, or when load shedding when enqueing messages.
-
Fix ordering bug in webhook subscribe topic authentication and topic rewrites (#823)
-
Fix issue when terminating the
vmq_server
application (#828). -
Make VerneMQ build on SmartOS / Illumos / Solaris.
-
Ensure strings passed from Lua to the logger are escaped (#864).
-
Handle Mongo Date / ISODate datatype properly in Lua / vmq_diversity (#857).
-
Optimize subscribe/unsubscribe operations for large fanout cases.
-
Allow non-standard MQTT version 131 (MQTT 3.1 bridge) by default (this was accidentally changed in VerneMQ 1.4.0).
-
Improve error messages returned by the
vmq_diversity
plugin so it's easier to understand and debug authentication and authorisation issues. -
Improve performance for looking up queue processes.
-
Make VerneMQ run on Erlang/OTP 21. To do this we had to upgrade
rebar3
to version 3.6.0 andlager
to version 3.6.3. We also removed theeper
dependency and suppressed some new warnings due toerlang:get_stacktrace/0
having been deprecated in Erlang/OTP 21. -
The bridge plugin (
vmq_bridge
) now has an option to queue outgoing message in case the network connection disappears. This is configurable with themax_outgoing_buffered_messages
setting on a bridge. In addition the bridge now has a simple cli interface undervmq-admin bridge show
. This work was kindly sponsored by Diacare-Soft(http://diacare-soft.ru). -
Fix multiple message retry issues in the MQTT client used by vmq_bridge.
-
Fix issue where the message ordering wasn't preserved after a client reconnect.
-
Add experimental
vmq_swc
plugin that provides an alternative to the existingvmq_plumtree
for metadata storage and replication. One must compile VerneMQ withmake swc
to generate a release containing thevmq_swc
plugin. -
Remove unused
vmq_bridge
sslcapath
config option. This was never used internally. -
Upgrade
lager
to version 3.6.8 to fix build issue on Mac OS X Mojave.
- Fix issue in the bridge preventing it from reconnecting after a connection timeout (#726).
- Fix issue with self referential cluster leaves and cluster readiness probing in general (#717).
- Add preparations for MQTTv5:
- Make it possible to extend the cluster internal message format when adding new features while ensuring back- and forwards compatibility.
- Refactor the leveldb message store code to support versioned indexes as well as messages.
- Refactor the retained message store to support versions.
- Add support for subscriber data with subscription options.
- Fix issue when validating a shared subscription topic (#756).
- Fix issue with retried PUBREC frames (#750).
- Make it possible to suppress the Last Will and Testament message if an
existing session already exists for the connecting client. This is
configurable via the hidden setting
suppress_lwt_on_session_takeover
in the VerneMQ configuration file. The default isoff
. - Fix issue with PUBREL frames retried after client reconnects (#762).
- Refactor and cleanup retry mechanism.
- Warn about deprecated setting
message_size_limit
only when it has been defined. - Change build script so nightly packages will have names on the form
vernemq_1.4.0-10-gde1b1f5-1_amd64.deb
, wherede1b1f5
refers to the commit hash from which the package was built and10
is the number of commits since the latest tag. This makes it much easier to reason about nightly builds and the features they contain. - Refactor and generalize the metrics systems to allow labelling metrics coming
from different sources in order to differentiate them in the various exporters
and the command line. Labels are added to the Prometheus exporter by
default. To enable generating additional metrics from the labels in the
graphite exporter the hidden setting
graphite_include_labels
has to be set toon
in thevernemq.conf
file. Labels are not exposed in the $SYS metrics. - Ensure the
vmq_bridge
is properly restarted after a crash (#785). - Fix issue where not calling
http.body(ref)
in a Lua script would not return the underlyning HTTP connection to the connection pool. This now happens automatically, irregardless of the user callinghttp.body(ref)
or not (#588).
- Fix for OSX compilation issue on
vmq_passwd
due to openssl headers not found. - Refactoring to further abstract and generalize the Plumtree metadata layer.
- Small refactoring moving the calling of plugin hooks into the fsm code. This
is a preparation for MQTTv5. Note, this change has an impact on the
vmq_reg:direct_plugin_exports/1
function. Even though this function is internal and as such not guaranteed to be stable, we know some plugin developers use it and we therefore wanted to describe the changes. It works as before, except that the hooksauth_on_subscribe
,on_subscribe
oron_unsuscribe
hooks will no longer be called when using the exported functions. - Corrected a few spelling errors in the
vmq-admin
help commands. - Added two new hidden configuration parameters for
vmq_webhooks
.vmq_webhooks
uses the hackney HTTP client for HTTP requests. It starts its own hackney pool with a default connection pool of 100, and a connection timeout of 60000 milliseconds. This connection pool con now be configurable using the settingsvmq_webhooks.pool_max_connections
andvmq_webhooks.pool_timeout
(#621). - Add a first version of a VerneMQ status page showing node and cluster information. This web-page is by default available on http://localhost:8888/status. The status page is implemented using jQuery, Twitter Bootstrap und mustache.js. As this is the first version the page should be considered experimental and will likely be changed in future releases.
- Bugfix: Handle return values correctly when enqueuing messages to offline shared subscribers (#625).
- Bugfix: Fix issue preventing messages delivered to a subscriber group from being delivered to the online group members first before attempting delivery to offline queues (#618).
- Fix some Dialyzer issues.
- Reduce replication load during a netsplit by making sure data is not attempted to be replicated to unreachable nodes.
- Bugfix: Fix issue causing some
session show
options (peer_host
,peer_port
) to not work for websocket clients (#542). - Bugfix: Fix routing table initialization issue after restarting a node preventing shared subscriptions on a remote node from being included in the routing table (#595).
- Bugfix: Fix race condition when fetching data from the internal query subsystem.
- Fix build issue on Raspberry PI (
make rpi-32
). - Make it possible to specify which protocol versions are allowed on an MQTT
listener. By default the protocol versions allowed are versions 3 and 4 (MQTT
v3.1 and v3.1.1 respectively). To set the allowed protocol versions one can
use
listener.tcp.allowed_protocol_versions = [3,4]
on the transport level or for a specific listener usinglistener.tcp.specific_listener.allowed_protocol_versions
. - Fix bug causing an exception to be thrown when
vmq-admin cluster leave
is used with a timout value less than 5 seconds (#642). - Small refactoring enabling to store versioned message store values. This is a preparation for MQTTv5.
- Bugfix: Fix a bug that prevented user plugins with an explicit path to be loaded.
- Fix issue with new rebar3 upstream plugin version (the port-compiler) which made builds fail by pegging it to an older version (1.8.0).
- Add
xmerl
andinets
from the Erlang standard library to the release in order to allow plugin developers to have these libraries available. - Bugfix: Fix typo (
graphie_api_key
->graphite_api_key
) preventing the graphite api key from being set in thevernemq.conf
file. - Bugfix: WebHooks Plugin. Close the Hackney CRef so that the socket is given back to the Hackney pool, for the case of non-200 HTTP OK status codes.
- Bugfix: fix bug where queries with mountpoint and client-ids would return no entries causing disconnecting a client with a specific mountpoint to fail. Also ensure to use the default mountpoint if no mountpoint was passed. (#714).
- Upgraded vernemq_dev to include the
disconnect_by_subscriber_id/2
API.
-
Add
proxy_protocol_use_cn_as_username
feature which forproxy_protocol
enabled listeners enable or disable using the common name forwarded by the PROXY protocol instead of the MQTT username. To ensure backward compatibility this setting is enabled by default. This feature was kindly contributed by SoftAtHome (https://softathome.com/). -
New
vmq-admin
command to forcefully disconnect and cleanup sessions. -
Fix issue preventing ssl settings being inheritable on the listener level (#583).
-
Fix issue where enqueuing data to a queue on a remote cluster node could cause the calling process to be blocked for a long time in case of the remote cluster node being overloaded or if a net-split has occurred.
This issue can occur while delivering a shared subscriber message to a remote subscriber in the cluster (blocking the publishing client) or when migrating queue data to to another node in the cluster. In the case of shared subscribers a new (hidden) configuration parameter (
shared_subscription_timeout_action
) has been added which decides which action to take if enqueuing on a remote note times out waiting for the receiving node to acknowledge the message. The possibilities are to eitherignore
the timeout orrequeue
the message. Ignoring the timeout can potentially lead to losing the message if the message was still in flight between the two nodes and the connection was lost due to a net-split. Requeueing may lead to the same message being delivered twice if the original client received the message, but the acknowledgement was lost. This fix was kindly contributed by SoftAtHome (https://softathome.com/). -
Fix typo in configuration name
plumtree.outstandind_limit
should beplumtree.outstanding_limit
. -
Make the
vmq_cluster_node
processes able to handle system messages to make it easier inspect the process behaviour at run-time. -
Fix bug preventing
use_identity_as_username
from working on WSS sockets (#563).
- The shared subscriber feature is no longer marked as in beta.
- Strengthen check verifying if a remote node is available or not. The new check
verifies that a data connection to the remote node has been established, in
addition to the current check which verifies that a specific process is running
on the remote node. This new check will make it more visible if, for instance,
the IP an port configured via
listener.vmq.clustering
is not reachable or a listener was not able to be started. - Fix issue preventing the proxy_protocol setting (
listener.tcp.proxy_protocol = on
) being inherited by specific listeners (#516). - Remove shared subscriptions backwards compatibility code. Introducing shared subscriptions changed some internals and code ensuring backwards compatibility had to be written. This code has existed since VerneMQ 1.0.0rc1 and anyone needing to upgrade a running cluster from older versions than 1.0.0rc1 of VerneMQ will have to first do an intermediate upgrade to a version containing the compatibility code.
- Remove time and randomness related dependencies for backwards compatibility for OTP 17. These are no longer required as OTP 17 support was removed before VerneMQ 1.0.
- Minor optimizations.
- Fix issue in the queue initialization introduced in VerneMQ 1.2.2 which meant offline messages were not being read into the queue process after a node restart. Added tests to prevent this issue from occurring again.
- Fix crash in cleanup of the
vmq_webhooks
plugin when the plugin is being stopped or the broker is shutting down (#556). - Fix incorrect format strings in
vmq_reg_sup
log statements. - Add missing argument to HTTP module configuration log statement.
- Do not resolve host names when including peer host in an
vmq-admin session show
query as this can lead to a timeout. - Remove superfluous warnings for ip and port when using the proxy protocol.
- Fixed a number issues when filtering
vmq-admin session show
results. Notemsg_ref
filtering is currently broken. - Fix node startup routine to cleanup stale subscriber data.
- Do not log getting a client peername failed because the client already disconnected.
- Upgrade MongoDB driver.
- Prevent unbounded memory growth in a netsplit situation by putting an explicit limit (250K) on the number of outstanding plumtree-msgs in plumtree. If this limit is exceeded, delayed replication will be handled by active anti-entropy.
- Handle uncaught error type in the
vmq_ql_query
handler. - Make sure the
peer_host
can always be retrieved via the HTTP API. It was returned as an erlang tuple which caused the conversion to JSON to fail. - Fix issue causing too few results to be returned from
vmq-admin session show
when used with filter options. This could happen when terms included in the filters also existed as erlang atoms. - Plugin workflow improvements: move plugin development specific things into
vernemq_dev
. - Fix error in the HTTP API interface. The alias
/api/v1/sessions
mapped tosession list
which is deprecated forsession show
. - Fix
max-age
parse issue invmq_webhooks
(#527).
- Fix retain msg server race condition preventing some messages from being persisted and replicated to other nodes in the cluster (#507).
- Log when a client is disconnected and multiple sessions are not allowed.
- Fix tracer
mountpoint
parameter bug. - Make it possible to add/inject HTTP API keys.
- Add Erlang VM memory usage stats to the exposed metrics.
- Fix bug with
max_message_size
andmessage_size_limit
only one of these should exist andmessage_size_limit
has now been deprecated and no longer has any effect.max_message_size
should be used instead. - Fix bug in
vmq_diversity
where the MongoDB client pool would not reconnect after a MongDB restart (#475). - Fix bug in
vmq_diversity
where the PostgreSQL client pool would not reconnect after a PostgreSQL restart (same as #475).
- Fix bug preventing persistent session expiration
(
persistent_client_expiration
invernemq.conf
) from being executed. - Make
vmq-admin session show
more robust when sessions are overloaded by limiting the time allowed to query each session. The default query timeout is 100ms, but can be overriden using--rowtimeout=<TimeoutInMilliseconds>
. - Add support for Erlang/OTP20.
- Improve tracer usage text.
- Fix
vmq_diversity
memcached issue (#460). - Fix RPM package issue preventing a clean upgrade.
- Fix code path bug related to upgrading plugins.
- Update
plumtree
with a fix to make the plumtree mailbox traversal measurement more accurate and therefore less spammy.
- Add more detail to the client expiration log message.
- Safeguard Lua function calls
- Fix bug where websocket connections were incorrectly terminated (#387).
- Fix breakage of event hook chain in
vmq_passwd
(#396). - Lua script balancing for improved Lua hook performance.
- Add descriptions to all available metrics via
vmq-admin metrics show --with-descriptions
. - Add Prometheus HELP lines for every metric.
- Add a log message (info level) indicating that connectivity to a remote node has been restored.
- Minor changes to the formatting of the tracer output.
- Purge node-local clean session true subscriber data when restarting VerneMQ as these would otherwise erroneously still exist.
- The metrics reporting sizes of retained messages
gauge.retain_memory
and routesgauge.router_memory
were incorrectly done in system word sizes instead of in bytes. This has been corrected. Further thegauge.retain_memory
metric now also includes temporary storage used before persisting messages to disk. - Fix bug causing hooks to be registered multiple times when reloading lua scripts (#348).
- Fix bug occurring when publishing across nodes where more than one subscriber are on one node and the publisher on another. In this case only one of the subscribers would receive the message.
- Fix formatting bug in the
vmq-admin trace
command. - Handle empty modifier list correctly in
vmq_webhooks
(#339). - Handle client_id and mountpoint modifiers correctly in
vmq_webhooks
(#332). - Fix vmq-admin session show when multiple filters are applied
- Fix bug where an aborted connection handshake caused queued messages not being migrated properly to remote node.
-
To make the
vmq-admin
tool more consistent, the following changes have been made:- vmq-admin script status -> vmq-admin script show
- vmq-admin session list -> vmq-admin session show
- vmq-admin cluster status -> vmq-admin cluster show
- vmq-admin webhooks status -> vmq-admin webhooks show
- vmq-admin webhooks cache stats -> vmq-admin webhooks cache show
- vmq-admin api create-key -> vmq-admin api-key create
- vmq-admin api delete-key -> vmq-admin api-key delete
- vmq-admin api list-keys -> vmq-admin api-key show
-
Fix a bug in
vmq-admin session list
triggered when running in a cluster. -
Add automatic CRL refreshing, configurable via the hidden
crl_refresh_interval
config option. Default is 60 seconds. -
Stop
all_till_ok
hook evaluation if a hook returns anything else thannext
. -
Add
vmq-admin trace
tracing sub-system. -
Improve help texts for
vmq-admin listener
command.
- Fix mongodb authentication problem due to a missing dependency.
- Fix wrong lua paths in generated packages.
- Fix connack counter metric
- Fix keepalive timer to fire earliest at 1.5 times the keepalive value
- Improve error reporting for wrong configuration parameters in vernemq.conf
- Fix various Dialyzer type errors
- Add out-of-the-box authentication and authorization support for Postgres,
MySQL, MongoDB, and Redis via
vmq_diversity
. - Erlang 17.x is no longer officially supported.
- New
rebar3
version (3.3.5), required to upgradenode_package
dependency. - The plugins
vmq_webhooks
andvmq_diversity
are now shipped as part of VerneMQ itself. - Packages are now built using Erlang/OTP 19.2.
-
Improved the
vmq-admin list session
command with a limit option controlling the returned number of sessions as well as the possibilty to customize which session data is returned. -
Change to plugin administration. To make VerneMQ configuration more consistent and simpler to configure it is now possible to configure plugins directly in the
vernemq.conf
file.-
The
lib/vmq_plugin.conf
file is obsolete and no longer in use. -
plugins are no longer persisted when enabled via the
vmq-admin
tool, but have to be added to thevernemq.conf
file manually like any other configuration. An example looks like:plugins.myplugin = on
plugins.myplugin.path = /path/to/plugin
Configuration specific settings are then configured like this:
myplugin.setting = ...
-
The above changes has the following impact on existing settings in the
vernemq.conf
file:-
The
acl_file
setting is replaced byvmq_acl.acl_file
. -
The
acl_reload_interval
setting is replaced byvmq_acl.acl_reload_interval
-
The
password_file
setting is replaced byvmq_passwd.password_file
-
The
password_reload_interval
setting is replaced byvmq_passwd.password_reload_interval
. -
The
bridge
prefix has been replaced by thevmq_bridge
prefix.Make sure to update the configuration file accordingly.
-
-
-
Add implementation of shared subscriptions as specified in the MQTTv5 draft spec. Currently two different distribution policies are supported:
prefer_local
which will, if possible, deliver the message to a random local subscriber in the shared subscription group, otherwise the message will be delivered to a random remote subscriber of the shared subscription (if any).random
will distribute the messages random between all members of the shared supscription regardless if they are local or are subscribed via another node.local_only
will consider only node-local members of the subscriber group for delivery.
A message published to a shared subscription will first be delivered to online members of the shared subscription if any exist, failing that, the message will be delivered to an offline subscriber of the shared subscription. Due to the fact that the MQTTv5 spec is not yet final, this feature is marked as BETA.
NOTE: To upgrade a live cluster all nodes must already be running 0.15.3 or newer! This feature is incompatible with older releases.
-
Use of specific routing tables for non-wildcard topics. This improvement results in faster routing table lookups for non-wildcard subscriptions, and reduces overall memory consumption of the routing tables.
- Minor fixes and improvements.
- Add support for caching ACLs in lua scripts.
- Add support for bcrypt password hashes.
- Fix a bug in the retain messages fastpath code.
- Add the first version of the VerneMQ HTTP admin API (BETA).
- Fix a nodename change bug, fix contributed by @dcy.
- Fix an oversight preventing the client id from being overridable in
auth_on_register
.
- Fix a bug in the MQTT parser.
- Make compatible with the OpenSSL 1.1 API.
- Fix a queue migration corner case with racing subscribers
The VerneMQ dependencies vmq_server, vmq_acl, vmq_passwd, vmq_plugin, vmq_commons were moved to the erlio/vernemq apps folder. The repos are kept around for some time, and the issue trackers are moved to erlio/vernemq.
This release contains some backward-incompatibilities with 0.14.2. Get in touch if professional support is required for upgrading.
- BC-break: New Subscriber format, the old format gets automatically translated to the new format. However in a clustered setting nodes running the old format can not deal with the new format once those records are replicated.
- BC-break: More control over behaviour during netsplits:
New configuration parameters
allow_register_during_netsplit
allow_subscribe_during_netsplit
allow_unsubscribe_during_netsplit
allow_publish_during_netsplit
obsolete thetrade_consistency
configuration parameter. Which was often misused together withallow_multiple_sessions
in order to reach high availability. With the new configuration parameters the behaviour has to be explicitly configured. The oldtrade_consistency
configuration parameter is not supported anymore.
- BC-break:
on_offline_message/1
hook replaced withon_offline_message/5
The new hook includes the SubscriberId, QoS, Topic, Payload, and RetainFlag. - Stability and performance improvements: In order to protect the broker from reconnect storms multiple processes implement the gen_server2 behaviour instead of gen_server. Moreover the queue subscriber is now load balanced to be able to cope with massive amounts of firing dead links. Improved retain performance for subscriptions without wildcards
- Support for PROXY protocol
- TLS certificate verification depth can be specified in vernemq.conf
- The
max_message_size
configuration now acts on the MQTT frame level instead of the MQTT payload. This has the benefit of preventing that too big messages reach the MQTT state machine. However this now includes every MQTT frame not only the publish payload. - TCP Socket options can now be configured on a session level by returning
a
{tcp_opts, list()}
modifier in theauth_on_register
hook.
- New configuration format enables to use host strings instead of only IP addresses.
- Protect broadcast module from overloading
-
SSL/TLS Refactorings: GCM cipher suites are now supported. The
support_elliptic_curve
option was removed. If your setup doesn't support the elliptic curve cipher suites you may do nothing, as the Erlang ssl application automatically ignores the unsupported cipher suites. As an alternative you may specify the supported ciphers in thevernemq.conf
file. -
VerneMQ is now compatible with Erlang 19: As of the upcoming VerneMQ 0.15.0 we'll switch to Erlang 19 as the default version. Backward compatibility with Erlang 17 is likely to be removed.
-
Bug fix in queue migration
- BC-break: Refactoring of queue migration mechanism: This change requires that all cluster nodes use 0.14.1. The refactoring fixes multiple issues with multiple clients that use the same client id when connecting to different cluster nodes.
- Metrics overflow fix
- Fix shutdown of plugins:
Disabling of application plugins that also dynamically register module plugins
(like
vmq_diversity
orvmq_webhooks
) will properly cleanup thevmq_plugin.conf
.
- RaspberryPI 32bit profile
Use
make rpi32
target to build a VerneMQ release for RaspberryPI
-
Major refactoring of the metrics subsystem: Graphite and Systree metric reporters are now part of vmq_server and don't rely on exometer_core anymore. Instead of exometer_core a more efficient Erlang NIF based counter implementation is used (mzmetrics). As a result of superseding exometer_core many new metrics are available now, and most of the old ones don't exist anymore. With the removal of the exometer_core dependency we gave up the native support for SNMP (in vmq_snmp). We introduce a Prometheus metrics HTTP handler, which enables the great open source monitoring solution Prometheus (with a rich set of reporters, including SNMP) to scrape metrics from VerneMQ. Please check the documentation and adjust your monitoring solutions.
-
LevelDB message store performance improvement for QoS 1 and 2 message
-
Fix a memory leak in message retry mechanism
-
Fix a message leak in the LevelDB message store: Since VerneMQ 0.12.5p4 the leaked messages were detected on broker start and automatically garbage collected.
-
Simplify changing the nodename in case of a single node cluster
-
MQTT conformance, session present MQTT-3.2.2-2
-
MQTT conformance, will message must not be published if client sends 'DISCONNECT' message: MQTT-3.1.2-8
-
MQTT conformance, unauthorized publishes are positively acknowledged: MQTT-3.3.5-2
-
MQTT conformance, pass empty retained message on to subscribers: MQTT-3.3.1-11
-
Several minor bugs fixes
-
Lines starting with '#' are treated as comments
-
Added mountpoint pattern '%m' support for pattern matches
-
Minor bug fixes
- Minor bug fixes
- Not included anymore, systree metrics are now reported by
vmq_server
: Make sure to check the new configuration for systree. Please check the documentation and adjust your monitoring solutions.
- Not included anymore, graphite metrics are now reported by
vmq_server
: Make sure to check the new configuration for graphite. Please check the documentation and adjust your monitoring solutions.
- Not included anymore, SNMP isn't natively supported anymore:
vmq_server
has a new HTTP metrics handler that exports all metrics in the Prometheus format. Prometheus is an open source monitoring solution that has multiple options for reporting metrics, including SNMP.
- Minor bug fixes
- Binary packages are now using Erlang OTP 18.3, this requires a recompilation of your custom Erlang plugins.
-
Major refactoring that enables a more scalable and robust migration of online and offline queues.
-
Use of new Rebar3 including the use of hex packages whenever possible.
-
Support for Erlang 18.
-
Reworked interal publish interface for plugins that require more control over the way VerneMQ publishes their messages (e.g. setting the dup & retain flags). Plugins which used the old 'PublishFun/2' MUST use the new 'PublishFun/3' which is specified as follows:
`fun([Word:binary()|_] = Topic, Payload:binary(), Opts:map()) -> ok | {error, Reason}`
-
Several minor bug fixes.
-
Several third-party dependencies have been upgraded.
-
Use of new Rebar3 and use of hex packages in all our own dependencies.
-
The docs are now hosted on our website http://verne.mq/docs.
This release increases overall stability and performance. Memory usage for the subscriber store has been reduced by 40-50%. To achieve this, we had to change the main topic structure, and unfortunately this breaks backward compatibility for the stored offline data as well as the plugin system.
.. warning::
Make sure to delete the old message store folder ``/var/lib/vernemq/msgstore``
and the old metadata folder ``/var/lib/vernemq/meta``.
If you have implemented your own plugins, make sure to adapt to the new topic
format. (Reach out to us if you require assistance).
-
Major refactoring of the MQTT state machine: We reduced the number of processes per connected client and MQTT state machine. This leads to less gen_server calls, fewer inter-process messages and a reduced scheduler load. Per client backpressure can be applied in a more direct way.
-
New topic format: Topics are now essentially list of binaries e.g. the topic
hello/+/world
gets parsed to a[<<"hello">>, <<"+">>, <<"world">>]
. Therefore every API that used a topic as an argument had to be changed. -
Improved cluster leave, and queue migration: This allows an operator to make a node gracefully leave the cluster.
- If the node is still online and part of the cluster a two step approach is used. During the first step, the node stops accepting new connections, but keeps serving the existing ones. In the second step it actively kills the online sessions, and (if possible) migrates their queue contents to the other cluster nodes. Once all the queues are migrated, the node leaves the cluster and terminates itself.
- If the node is already offline/crashed during a cluster leave call, the old subscriptions on that node are remapped to the other cluster nodes. VerneMQ gets the information to do this mapping from the Plumtree Metadata store. No offline messages can be copied in this case.
-
New hook,
on_deliver/4
: Every message that gets delivered passes this hook, which allows a plugin to log the message and much more if needed (change payload and topic at delivery etc) -
New hook,
on_offline_message/1
: If a client has been connected withclean_session=false
every message that gets offline-queued triggers this hook. This would be the entrypoint to call mobile push notification services. -
New hook,
on_client_wakeup/1
: When an authenticated client attaches to its queue this hook is triggered. -
New hook,
on_client_offline/1
: When a client withclean_session=false
disconnects this hook is triggered. -
New hook,
on_client_gone/1
: When a client withclean_session=true
disconnects this hook is triggered. -
No RPC calls anymore during registration and queue migration flows.
-
Many small bug fixes and improvements.
-
New topic parser and validator
-
New shared behaviours for the new hooks
- Adapt to use the new topic format
- Adapt to use the new topic format
- Adapt to use the new topic format
The queuing mechanism got a major refactoring. Prior to this version the offline messages were stored in a global in-memory table backed by the leveldb based message store. Although performance was quite ok using this approach, we were lacking of flexibility regarding queue control. E.g. it wasn't possible to limit the maximum number of offline messages per client in a straightforward way. This and much more is now possible. Unfortunately this breaks backward compatibility for the stored offline data.
.. warning::
Make sure to delete (or backup) the old message store folder
``/var/lib/vernemq/msgstore`` and the old metadata folder ``/var/lib/vernemq/meta``.
We also updated the format of the exposed client metrics. Make sure
to adjust your monitoring setup.
-
Major refactoring for Queuing Mechanism: Prior to this version the offline messages were stored in a global ETS bag table, which was backed by the LevelDB powered message store. The ETS table served as an index to have a fast lookup for offline messages. Moreover this table was also used to keep the message references. All of this changed. Before every client session was load protected by an Erlang process that acted as a queue. However, once the client has disconnected this process had to terminate. Now, this queue process will stay alive as long as the session hasn't expired and stores all the references to the offline messages. This simplifies a lot. Namely the routing doesn't have to distinguish between online and offline clients anymore, limits can be applied on a per client/queue basis, gained more flexibility to deal with multiple sessions.
-
Major refactoring for Message Store The current message store relies only on LevelDB and no intermediate ETS tables are used for caching/indexing anymore. This improves overall memory overhead and scalability. However this breaks backward compatibility and requires to delete the message store folder.
-
Changed Supervisor Structure for Queue Processes The supervisor structure for the queue processes changed in a way that enables much better performance regarding setup and teardown of the queue processes.
-
Improved Message Reference Generation Performance
-
Upgraded to newest verison of Plumtree
-
Upgraded to Lager 3.0.1 (required to pretty print maps in log messages)
-
Many smaller fixes and cleanups
-
Better error messages in case of parsing errors
-
Fixed a parser bug with very small TCP segments
We switched to the rebar3 toolchain for building VerneMQ, involving quite some
changes in many of our direct dependencies (vmq_*
). Several bug fixes and
performance improvements. Unfortunately some of the changes required some backward
imcompatibilites:
.. warning::
Make sure to delete (or backup) the old subscriber data directory
``/var/lib/vernemq/meta/lvldb_cluster_meta`` as the contained data format isn't
compatible with the one found in ``0.10.0``. Durable sessions (``clean_session=false``)
will be lost, and the clients are forced to resubscribe.
Although the offline messages for these sessions aren't necessary lost, an ugly
workaround is required. Therefore it's recommended to also delete the message store
folder ``/var/lib/vernemq/msgstore``.
If you were running a clustered setup, make sure to revisit the clustering
documentation as the specific ``listener.vmq.clustering`` configuration is needed
inside ``vernemq.conf`` to enable inter-node communicaton.
We updated the format of the exposed metrics. Make sure to adjust your monitoring setup.
-
Changed application statistics: Use of more efficient counters. This changes the format of the values obtained through the various monitoring plugins. Added new system metrics for monitoring overall system health. Check the updated docs.
-
Bypass Erlang distribution for all distributed MQTT Publish messages: use of distinct TCP connections to distribute MQTT messages to other cluster nodes. This requires to configure a specific IP / port in the vernemq.conf. Check the updated docs.
-
Use of more efficient key/val layout within the subscriber store: This allows us to achieve higher routing performance as well as keeping less data in memory. On the other hand the old subscriber store data (not the message store) isn't compatible with the new one. Since VerneMQ is still pretty young we won't provide a migration script. Let us know if this is a problem for you and we might find a solution for this. Removing the '/var/lib/vernemq/meta/lvldb_cluster_meta' folder is necessary to successfully start a VerneMQ node.
-
Improve the fast path for QoS 0 messages: Use of non-blocking enqueue operation for QoS 0 messages.
-
Performance improvements by reducing the use of timers throughout the stack: Counters are now incrementally published, this allowed us to remove a timer/connection that was triggered every second. This might lead to accuracy errors if sessions process a very low volume of messages. Timers for process hibernation are removed since process hibernation isn't really needed at this point. Moreover we got rid of the CPU based automatic throttling mechanism which used timers to delay the accepting of new TCP packets.
-
Improved CLI handling: improved 'cluster leave' command and better help text.
-
Fixed several bugs found via dialyzer
-
Multiple conformance bug fixes: Topic and subscription validation
-
Improved generic gen_emqtt client
-
Added correct bridge protocol version
-
Fixed bugs found via dialyzer
-
Merged updated SNMP reporter from feuerlabs/exometer
-
Cleanup of unused OTP mibs (the OTP metrics are now directly exposed by vmq_server)
-
Minor bug fixed related to dynamically loading plugins
-
Switch to rebar3 (this includes plugins following the rebar3 structure)