Missing _monitoring system bucket

[qdel]

Steps to reproduce:

$> influx bucket list
ID                      Name            Retention       Shard group duration    Organization ID         Schema Type
1f0b7f93eee31365        _tasks          72h0m0s         24h0m0s                 0fb9ddd02e0b86db        implicit
5ac999f5d8a9894a        batterystats    infinite        168h0m0s                0fb9ddd02e0b86db        implicit
23e6a8a6b5544aeb        default         168h0m0s        24h0m0s                 0fb9ddd02e0b86db        implicit
1e44de25ae87b286        docker          720h0m0s        24h0m0s                 0fb9ddd02e0b86db        implicit
84b910126dacae33        linky           infinite        168h0m0s                0fb9ddd02e0b86db        implicit
42c4db6643210999        nginx           720h0m0s        24h0m0s                 0fb9ddd02e0b86db        implicit
72c711e492786cd0        telegraf        2160h0m0s       24h0m0s                 0fb9ddd02e0b86db        implicit
f8241b0bbbb34483        telegraf_1m     infinite        168h0m0s                0fb9ddd02e0b86db        implicit
fbc48ead3cda9628        veolia          infinite        168h0m0s                0fb9ddd02e0b86db        implicit

Expected behaviour: _monitoring should be present

Actual behaviour: It is not

Environment info:

Very old instance, using influxdb v1. All data migration made.

Config:

Started using docker compose. But config is really simple:

  influxdb:
    image: influxdb:alpine
    ports:
      - "8086:8086"
    volumes:
      - /opt/influxdb2:/var/lib/influxdb2
    #environment:
    #  - INFLUXD_LOG_LEVEL=error
    networks:
      - super
    restart: unless-stopped
    cpu_shares: 1800

Logs:

No relevant logs found.

[philjb]

I don't believe v1 OSS supports _monitoring system bucket (as in I believe _monitoring was added in v2).

[qdel]

Indeed. I think that albeit all migrations where ok, the _monitoring bucket did not create itself.

Nearly everything is working fine. But when creating a new bucket, when I want to write new measurement in the new bucket I got errors about _monitoring bucket.

All in place systems have no troubles (maybe why I never saw the glitch before)

[qdel]

Bump on this. Is there any way to re-create system bucket ?

[philjb]

Have you tried trying to create _monitoring directly/manually? It might let you since it isn't present, but I'm not sure what it allows in this situation since you've gotten into an unexpected state.

Alternatively, you could make a backup and import into a new instance of oss 2.x which correctly has the monitoring bucket already. Probably not ideal.

[qdel]

How can I do it? The UI does not allow it and CLI documentation clearly specify that bucket should not be created with _ as first character.

Could be possible but time and logistic consuming.

It would be easier if influxdb recreate those if needed.

Le mer. 16 oct. 2024, 17:25, Phil Bracikowski @.***> a écrit :

Have you tried trying to create _monitoring directly/manually? It might let you since it isn't present, but I'm not sure what it allows in this situation since you've gotten into an unexpected state.

Alternatively, you could make a backup and import into a new instance of oss 2.x which correctly has the monitoring bucket already. Probably not ideal.

— Reply to this email directly, view it on GitHub https://github.com/influxdata/influxdb/issues/24918#issuecomment-2417165057, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAJIGL75BD27Q7CBJOR2YMLZ32AMZAVCNFSM6AAAAABQBHLS4GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIMJXGE3DKMBVG4 . You are receiving this because you authored the thread.Message ID: @.***>