1
83%
main: 83%

Ran 27 Aug 2024 05:37AM UTC

Files 56

Run time 2s

Badge

Embed ▾

Committed 26 Aug 2024 09:54PM UTC coverage: 85.153% (-0.09%) from 85.242%

Job # 10571820421.1

Build Type

push

github

Committed by

web-flow

Commit Message

Don't return `seq=0, duplicate=true` when processing duplicate messages asynchronously (#5827)

This issue was reported by Marco and Antithesis.

`seq=0, duplicate=true` would be returned if two duplicate messages
would be processed at the same time. The first will propagate, which
will take some time for a clustered stream, but then the second message
gets processed and rejected with `seq=0, duplicate=true` (without
`error`) which would be invalid.

One option to fix this is timing out the second publish and wait for the
original message to finish processing.
That fixes this issue, and the user would just retry the request and get
the valid sequence once the other message has gone through.

However, it could decrease message throughput when this is hit. An
alternative could be to throw an error, but if the user would always
retry if `err != nil` then they'd be "spamming" the same request until
the original message is fully propagated.

Would like to hear your thoughts :) 
Maybe this would be more suitable for 2.11.x as well?

Signed-off-by: Maurice van Veen <github@mauricevanveen.com>

Coverage Stats

66302 of 77862 relevant lines covered (85.15%)

1214568.37 hits per line

nats-io / nats-server / 10571820421 / 1
83%
main: 83%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 10571820421.1

nats-io / nats-server / 10571820421 / 1 83% main: 83%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on job 10571820421.1

nats-io / nats-server / 10571820421 / 1
83%
main: 83%

README BADGES
x