-
barfieldgreetings amigos! I have a headnode that I had to swap out to a different chassis. Trying to figure out why rabbitmq0/mdata:execute svc will not start is giving me a splitting headache
-
barfieldanyone have any guidance on this? I hate erlang
-
barfieldsvcs -xv says registrar is not running, but svccfg export mdata:execute says that it depends on registrar WTF
-
barfieldThey depend on each other lol
-
barfieldalso there is this
-
barfieldError: ur client cannot connect to AMQP broker.
-
barfieldError: Results are only for local (headnode) services.
-
barfieldoh weird I just noticed that there are two rabbit mq instances listed in sdcadm check-health
-
barfieldone with a UUID of all 0's
-
barfield00000000-0000-0000-0000-000000000000 rabbitmq headnode rabbitmq0 false
-
barfield35a3d51f-2272-4e32-8bd5-6aea2411dc8c rabbitmq headnode rabbitmq0 false
-
barfieldchecking sapi records
-
barfieldcnapi keeps dying because it can't rabbitmq
-
neuroservequestion - which one is the right one?
-
bahamatAs with all compute node chassis swaps, the old UUID is not removed from Triton. You have to do that manually.
-
barfieldIn which place? I deleted it from cnapi and restarted cn-agent
-
bahamatYou'll probably need to clean up nics and agents separately.
-
barfieldI cleaned up all of the NICS already
-
barfieldcmon agent turned out to be the culprit for all of that junk in cnapi under the headnode that I asked you about not long ago. The junk was in the agents.cmon-agent.uuid field
-
barfieldI generated a new one and updated it
-
barfieldBut I am just not sure where else to update
-
bahamatWell sdcadm looks at sapi.