| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |\ |
|
| | |
| |
| |
| |
| | |
Conflicts:
scripts/rabbitmq-env.bat
|
| | |\
| | |
| | | |
Improve `partitions_SUITE`
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The testcase simulates a partial partition which should be detected by
the two nodes on each side of the partition. They should both decide
there is a partial partition and stop.
However, sometimes one of the node is faster than the other one: it sees
the partial partition and stops before the second node even sees the
partition. When this happens, the second node doesn't even bother to
stop (because the partition is full now). Thus the testcase fails
because it expects both nodes to stop.
This patch tries to sync partial partition checks so both A and B
decide there is a partial partition.
[#146911969]
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The testcase simulates a small delay between the times B sees A down
and C sees A down too. Because A is effectively down for both, B's
rabbit_node_monitor shouldn't see a partial partition.
When B sees A down first, it asks C about A. In the scenario the
testcase simulates, C tries to contact A, fails and reports to B that A
is down.
However, because of the way we simulate that delay, sometimes, C still
sees A when B asks about it, because the link between A and C isn't
blocked yet. When this happens, B decides there is a partial partition.
Therefore, the testcase fails because it tests that this shouldn't
happen.
To try to be closer to a real-world situation (where A would be really
down for B and C), we suspend C's rabbit_node_monitor until we block the
link between A and C. Thus B still asks C about A, but C only processes
the request when the link is indeed down.
[#146911969]
|
| | |/ |
|
| | | |
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Tests were failing because they required vm_memory_monitor process to be running
re #1259
[#145451399]
Signed-off-by: Gerhard Lazu <gerhard@rabbitmq.com>
|
| | |
| |
| |
| |
| |
| | |
[#145451399]
Signed-off-by: Gerhard Lazu <gerhard@rabbitmq.com>
|
| | |
| |
| |
| |
| |
| |
| |
| | |
Account for extra memory that was reported in other_ets
[#145451399]
Signed-off-by: Gerhard Lazu <gerhard@rabbitmq.com>
|
| | |
| |
| |
| |
| |
| | |
According to AIX, SunOS and BSD docs, ps command can have rss format,
so get_ps_memory should work.
We already requiring tasklist for wait command, so we can use it here.
|
| |/
|
|
|
|
|
|
|
|
| |
Memory reported by erlang:memory(total) is not supposed to
be equal to the total size of all pages mapped to the emulator,
according to http://erlang.org/doc/man/erlang.html#memory-0
erlang:memory(total) under-reports memory usage by around 20%
[#1223]
[#145451399]
|
| |\
| |
| | |
OTP-20 compatibility workaround for queue directory names. 3.6 only!
|
| | | |
|
| | |\
| |/
|/| |
|
| |\ \
| | |
| | | |
Improve documentation around new default for background garbage collection
|
| |/ /
| |
| |
| | |
background_gc_enabled to false to match the behavior of the Makefile.
|
| |\ \ |
|
| |/ /
| |
| |
| |
| | |
rabbitmq-server#1233
[#146160779]
|
| | | |
|
| |/
|
|
|
|
|
| |
OTP-20 has different binary format for atoms (since OTP-14337)
So term_to_binary generates different value.
testm_to_binary_compat now have a function to generate a binary
identical to pre-20 versions.
|
| | |
|
| | |
|
| |\
| |
| | |
Update sup_delayed_restart_SUITE for rabbitmq-common#202
|
| |/
|
|
| |
To accommodate changes made in rabbitmq/rabbitmq-common#201.
|
| |\
| |
| | |
Don't expand plugins that are already unpacked
|
| | | |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This prevents creating unnecessary files when running in embedded
mode, and in development mode it keeps code path clean and compatible
with different interactive tools.
Also some new error conditions are properly handled and logged:
- Failure to unpack an .ez-file
- Missing .app file in plugin directory
|
| |\ \
| | |
| | | |
Introduce a new config parameter, total_memory_available_override_value, to override total amount of memory available to the node
|
| | | | |
|
| | |/
| |
| |
| |
| |
| |
| | |
memory
rabbitmq-server#1224
[#145450413]
|
| | |
| |
| |
| | |
rabbit_misc:json_encode can return a list or a binary.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| | |
This module is generic enough and is used by the Erlang client. Moving
it is the last step to resolve the interdependency between the Erlang
client and the broker: with this change, the client only depends on
rabbitmq-common!
[#118490793]
|
| | |
| |
| |
| |
| | |
Erlang 20.x warns if `export_all` is set in a module by default. We need
to turn the warning off.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Those are the functions to clear the read buffer cache in various
internal RabbitMQ processes. file_handle_cache still reads the
`fhc_read_buffering` and `fhc_write_buffering` rabbit application
variables, though.
Thus, this change almost removes a dependency of file_handle_cache on
the broker. Reading the configuration variable is acceptable for now so
this will allow us to move file_handle_cache to rabbitmq-common.
[#118490793]
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The function is renamed to get_memory_use/1 in the process, to match
other functions in the vm_memory_monitor module.
The file_handle_cache uses that function. By moving it to
vm_memory_monitor, we remove a dependency of file_handle_cache on a
rabbit_* module. This will allow us to move both file_handle_cache and
vm_memory_monitor to rabbitmq-common and help remove a dependency in the
Erlang client on the broker because the former calls
file_handle_cache:obtain/0.
To sum up the situation before this commit with a graph:
amqp_network_connection --[depends on]--> file_handle_cache
(amqp_client) (rabbit)
file_handle_cache --[depends on]--> rabbit_memory_monitor
(rabbit) (rabbit)
And the situation we want to reach after the move:
amqp_network_connection --[depends on]--> file_handle_cache
(amqp_client) (rabbit_common)
file_handle_cache --[depends on]--> vm_memory_monitor
(rabbit_common) (rabbit_common)
rabbit_memory_monitor --[depends on]--> vm_memory_monitor
(rabbit) (rabbit_common)
[#118490793]
|
| |
|
|
|
|
| |
See the corresponding commit in rabbitmq-common for an explanation.
[#144697185]
|
| | |
|
| |
|
|
| |
per discussion with @gerhard.
|
| |\
| |
| | |
Check rabbit is running after the boot process finishes
|
| | |
| |
| |
| |
| |
| | |
Signed-off-by: Gerhard Lazu <gerhard@rabbitmq.com>
[#145043957]
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Lower mnesia_table_loading_retry_timeout so test suite will be faster.
Add mnesia_table_loading_retry_limit so that we test retries.
Signed-off-by: Gerhard Lazu <gerhard@rabbitmq.com>
re #1214
[Finishes #145043957]
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
| |
Boot process can exit if boot fails. It can result in
rabbit:await_startup/0 returning ok & rabbitmqctl wait exiting with exit
status 0.
WIP, still need to add some tests.
Signed-off-by: Gerhard Lazu <gerhard@rabbitmq.com>
re #1214
[#145043957]
|
| |
|
|
| |
Signed-off-by: Gerhard Lazu <gerhard@rabbitmq.com>
|
| |\
| |
| | |
Test single slave restart after forget_cluster_nodes in offline mode
|
| | |
| |
| |
| |
| | |
While "Removee" is clear and to the point, some may conclude
that it is a typo (supposed to be "Remove").
|
| | |
| |
| |
| |
| | |
rabbitmq-server#1213
[#145084865]
|
| |\ \
| |/
|/|
| |
| | |
binarin/more-defaulting-for-plugins-expand-dir-stable
Use plugins expand dir with default fallback value
|
| |/
|
|
| |
In addition to #1134
|
