| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |\
| |
| | |
Pass all members to synchronous Ra operations
|
| | |\
| |/
|/| |
|
| |\ \
| | |
| | | |
Support client-driven refresh of expiring credentials
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
instead of disabling the cache entirely. Disabling the
cache has a significant double digit throughput drop effect
on super fast publishers, as demonstrated by
a number of PerfTest benchmarks.
Per discussion with @acogoluegnes, @kjnilsson.
References rabbitmq/rabbitmq-auth-backend-oauth2#28.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
Per discussion with @acogoluegnes.
While at it, correct a few type spec issues reported
by Dialyzer.
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Integration tests in rabbitmq_auth_backend_oauth2
cover more ground. In addition we would have to
introduce a dummy backend that supports state
updates which @acogoluegnes and I decided against.
So all the testing currently happens in the only
backend plugin that actually supports credential
expiration and state/secret updating.
|
| | | | |
|
| | |\ \ |
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | | |
Corrects a typo in 97f3f52.
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | | |
Spotted by @acogoluegnes.
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Channel processes use it for authz backend checks. The record
is identical to that of connection process (and originates from it), so
a simple cast to update the field is appropriate: any errors would
be handled and communicated back by the connection itself.
Per discussion with @acogoluegnes and @kjnilsson.
|
| | | | | |
|
| |\ \ \ \
| | | | |
| | | | | |
Provide timeout to add_member command
|
| | | |/ /
| |/| |
| | | |
| | | | |
[#166670827]
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
... in `undefined_channel_source` testcase. This testcase fails
transiently and it looks like a timing issue:
channel_source_SUITE > non_parallel_tests > undefined_channel_source
#1. {error,
{{badmatch,[]},
[{channel_source_SUITE,undefined_channel_source1,1,
[{file,"test/channel_source_SUITE.erl"},{line,138}]},
{rpc,'-handle_call_call/6-fun-0-',5,
[{file,"rpc.erl"},{line,197}]}]}}
|
| |\ \ \ \
| |/ / /
|/| | | |
Remove deprecation of force_event_refresh message
|
| |/ / /
| | |
| | |
| | |
| | |
| | | |
Part of rabbitmq/rabbitmq-management-agent#81
Also see rabbitmq/rabbitmq-management-agent#58
|
| | | | |
|
| | | | |
|
| |\ \ \
| | | |
| | | | |
Qq metrics tweak
|
| |\ \ \ \
| |_|/ /
|/| | | |
add test for cleanup of exclusive queues in case of restart or partition
|
| |/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
When using different than client-local queues exclusive ones can
be placed on a different host than the client. When there is a partition
or restart these queues are stuck because they are not cleaned up in
the nodes as the connection is still alive when the node goes offline,
however the queue is dead in the partition. These queues can not be
redeclared or consumed on 3.7.x.
Related: #2039
|
| | |/
| |
| |
| |
| |
| |
| | |
As this will make these more work reliably when leadership is in flux.
Also ensure that servers started and added after the initial declare
uses the same configuration.
|
| | |
| |
| |
| | |
To match it in unit with the rest of rabbit.
|
| |/ |
|
| |
|
|
| |
Addressing a TODO of that effect.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Plugins can provide a module to be called to extract extra information
at authentication time. The module name is based on the protocol used. Some
protocols have a space in their name, like "Web MQTT", so spaces in such
cases must be replaced by underscores to come up with a valid Erlang
module name.
[#166271318]
References #1767
|
| |\
| |
| | |
Workaround for IPv6 EPMD
|
| | |
| |
| |
| |
| | |
With this patch it is possible to pass variables to
prelaunch state.
|
| |\ \
| | |
| | | |
Make default worker pool size configurable
|
| |/ /
| |
| |
| | |
Part of rabbitmq/rabbitmq-server#2030.
|
| |\ \
| | |
| | | |
Schema information in mnesia for cli command
|
| | | |
| | |
| | |
| | | |
[#164550205]
|
| | | |
| | |
| | |
| | |
| | |
| | | |
Now, both modules export the same set of functions. This will help with
the backport to `v3.7.x` and thus the life of plugin developers who will
not have to have one copy of their plugin per RabbitMQ version.
|
| | | | |
|
| |\ \ \
| |/ /
|/| | |
Update log configuration tests for rabbitmq/rabbitmq-common#328
|
| |/ / |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Public IPv6 192.240.153.117 is managed by OVH US and has a PTR
117.153.240.192.in-addr.arp record that resolves to www.rabbitmq.com. It
would be nice to move to an IPv4 & IPv6 that are managed by our wider
org.
dig 117.153.240.192.in-addr.arpa PTR +short
www.rabbitmq.com.
whois 192.240.153.117 | grep -i Organiz
Organization: OVH US LLC (OUL-16)
cc @michaelklishin @trevormarshall
|
| |\ \
| | |
| | |
| | |
| | | |
rabbitmq/rabbitmq-server-1767-protocol-specific-ctx-in-authn-authz
Propagate protocol-specific context to authorization
|
| | |\ \
| |/ /
|/| |
| | | |
rabbitmq-server-1767-protocol-specific-ctx-in-authn-authz
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | | |
When one queue confirms a message and another queue failed
without confirming it, the message should be rejected.
Fixes some races between confirms and rejects.
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | | |
discard/4 confirms a message, which can cause races between confirms
and rejects.
|
| |\ \ \
| | | |
| | | | |
Add channel info for pending raft commands
|
| | | | |
| | | |
| | | |
| | | | |
[#165184345]
|
