Directory where this tool is downloaded should be readable by rabbitmq user. It’ll inject some code into a running broker instance. This code shouldn’t interfere with regular rabbitmq functioning. To purge it RabbitMQ needs to be restarted.
Establishing proper long-lived connection to RabbitMQ node requires
knowing nodename, erlang cookie and net ticktime. By default
Makefile tries to automatically get them using rabbitmqctl
, and
it should work most of the time.
But all this parameters can be manually set using make
parameters like this:
make shell N=nodename T=netticktime C=erlang-cookie
make shell
After running this command you’ll get an interactive erlang shell
to a running rabbitmq instance. Various helper functions from
user_default.erl
will be available here.
There are some tools to debug an issue when exchange/queue/bindings exists, but messages are not being routed to a queue.
First command perform tracing of message routing process (works better when HiPE is disabled)
make troute f=/path/to/produced/trace e=<exchange-name> r=<routing-key>
For cases when HiPE is enabled, running the second command is also necessary:
make tdebug d=/path/to/dump/directory/ e=<exchange-name> r=<routing-key>
Hopefully these 2 commands will collect enough information to debug the root cause of the routing issue.