Open tvojacek opened 1 year ago
Thanks for the report, did you have active mqtt traffic while doing this or see any other WRN logs?
Also, do you (mostly? exclusively?) use clean or "persistent" sessions in MQTT?
Problem lay in jetstream data of mqtt, these jetstreams has been migrated from 2.9.21 If I delete these data problem disappear. After recovery of buggy stream data problem repeat . streams.zip
Any updates on this one?
What version were you using?
nats 2.10.0,nats 2.10.1
What environment was the server running in?
docker nats:2.10.1-alpine3.18 or nats:2.10.0-alpine3.18
Is this defect reproducible?
start client connecting to mqtt publishing x messages every second start server: docker compose up -d create jetstream that capture message reload server (nats-server --signal reload or '$SYS.REQ.SERVER..RELOAD'
Given the capability you are leveraging, describe your expectation?
on 2.9.21 or if is not started, reload happen instantly.
Given the expectation, what is the defect you are observing?
There is 20second delay between Reloaded: MQTT consumer_inactive_threshold = 0s and Reloaded server configuration. During that period auth keys are not loaded. If client connect to get data from jetstream is get error:
nats | [1] 2023/10/06 11:02:27.495304 [ERR] 172.19.0.9:56102 - cid:34 - Publish Violation - Nkey "Uxxxxxx", Subject "$JS.MOCK00-1111-1111-0001.API.CONSUMER.MSG.NEXT.STX_DATA.FORWARDER"
After 20 seconds after
Reloaded server configuration
message problem with Nkey is resolved.