nats s view select topic T98 instead T03

[perestoronin]

Defect

Versions of nats-server and affected client libraries used:

nats-server version 2.9.0

Steps or code to reproduce the issue:

nats s view WAREHOUSE --context=stage_leaf --subject="WAREHOUSE.documents.T03" --since=4h

Expected result:

Only selected subject="WAREHOUSE.documents.T03

Actual result:

subject="WAREHOUSE.documents.T98

[derekcollison]

Could you share nats stream info for the stream?

[perestoronin]

Could you share nats stream info for the stream?

nats s info WAREHOUSE Information for Stream WAREHOUSE created 2022-03-29T14:16:10+03:00

Configuration:

         Subjects: WAREHOUSE.>
 Acknowledgements: true
        Retention: File - Limits
         Replicas: 1
   Discard Policy: Old
 Duplicate Window: 2m0s
Allows Msg Delete: true
     Allows Purge: true
   Allows Rollups: false
 Maximum Messages: unlimited
    Maximum Bytes: unlimited
      Maximum Age: unlimited

Maximum Message Size: unlimited Maximum Consumers: unlimited

State:

         Messages: 4,956,765
            Bytes: 5.6 GiB
         FirstSeq: 1 @ 2022-03-29T12:57:02 UTC
          LastSeq: 4,956,765 @ 2022-09-21T14:59:19 UTC
 Active Consumers: 364

[derekcollison]

Thanks, do you do any server restarts? Does ever NATS cli consumer that you create deliver incorrect messages if asked to be filtered or just some? Any timing specifics that might be helpful?

[kozlovic]

I am trying to reproduce, same topic name and subjects spaced filled with same amount of messages and random payload up to 1Kb and then creating a durable pull consumer and pulling on a specific subject and check that subject is correct. So far not able to reproduce. Going now to loop over all subjects and will report back.

[kozlovic]

I have been checking the 100 subjects that I have (WAREHOUSE.documents.T00 to WAREHOUSE.documents.T99) randomly, and the test passes. I am running from main branch, so will now try from actual v2.9.0 release to see if it makes a difference.

[ripienaar]

I have only seen this on streams that existed on the earlier version. Clustered and stand-alone. But can’t get a easy repro

[derekcollison]

@perestoronin was the stream created on a NATS server version < 2.9.0?

[kozlovic]

I have used 2.8.4 to generate the data, then run test to consume (randomly) from all subjects (in both the original version and v2.9.0), and so far no issue. Staring that the code at the moment...

[perestoronin]

@perestoronin was the stream created on a NATS server version < 2.9.0?

Yes, stream created in version 2.8.1 or 2.8.2, or more old ...

[derekcollison]

If the server is restarted again does the problem persist.

[derekcollison]

We think we now understand the issue. PR coming, thanks.

[kozlovic]

We think this is addressed with PR #3486.