Tenant pollution Orion v2.6.1, Orion upgrade will help us?

[objecsoft]

Bug description
"We are running two Orion Context Broker pods in a Kubernetes cluster with a multi-tenant setup, backed by approximately 40 MongoDB databases. We are observing tenant data pollution: a small percentage of ingested NGSIv2 entities are being persisted in an incorrect tenant's database, though the default tenant database remains empty."

• Orion version: 2.6.1 (this can be got with GET /version in the API port or using contextBroker --version)
  "orion": {
  "version": "2.6.1",
  "uptime": "0 d, 0 h, 17 m, 20 s",
  "git_hash": "ec20c8bcfd883d6d7214a28818c8310fb179bbcf",
  "compile_time": "Tue Mar 30 11:43:48 UTC 2021",
  "compiled_by": "root",
  "compiled_in": "06ed7e03551a",
  "release_date": "Tue Mar 30 11:43:48 UTC 2021",
  "doc": "https://fiware-orion.rtfd.io/en/2.6.1/",
  "libversions": {
  "boost": "1_53",
  "libcurl": "libcurl/7.29.0 NSS/3.53.1 zlib/1.2.7 libidn/1.28 libssh2/1.8.0",
  "libmicrohttpd": "0.9.70",
  "openssl": "1.0.2k",
  "rapidjson": "1.1.0",
  "mongodriver": "legacy-1.1.2"
  }
  }
• MongoDB version 4.2.7
• Env variables or CLI parameters
  "orion": {
  "log_level": "WARN",
  "mongo": {
  "auth": {
  "source": "admin"
  },
  "pool_size": "400",
  "db": "",
  "password": "",
  "user": "gip_orion_p",
  "replica_set": "repset_iot1_prod",
  "host": "mongo1:27017,mongo2:27017,mongo3:27017"
  },
  "notif": {
  "mode": "threadpool:10:200"
  },
  "stat": {
  "notif_queue": "TRUE",
  "counters": "TRUE"
  },
  "http_timeout": "3000",
  "trq_pool_size": "400",
  "ngsiv1_autocast": "TRUE",
  "multi_service": "TRUE",
  "foreground": "TRUE",
  "port": "80"
  }
  }
• Operating System or Docker Image used: we deployed 2 K8s pods for this old Orion version

How to reproduce it
HTTP POST to https://orion-app1.city.be/v2/op/update + tenantId as a http header eg. tenantId:city_awv
Body:
{
"actionType": "APPEND",
"entities": [
{
"id": "Vehicle:PoppyCar:595419bb-1d06-544e-bad3-ce278c8d61fc",
"type": "Vehicle",
"dataProvider": {
"type": "URL",
"value": "https://poppy.be"
}
},
{
"id": "Vehicle:PoppyCar:b3503e54-e26d-54d8-9f32-70cf49f93118",
"type": "Vehicle",
"dataProvider": {
"type": "URL",
"value": "https://poppy.be"
}
}
]
}
Expected behavior
No tenant pollution, all NGSIv2 entities are stored into the right MongoDB as explicitly mentioned with the tenantId header in the HTTP POST .../v2/op/update
?
Additional information
We will upgrade to the leatest greates Orion version 4.2, but will this solve the tenant pollution issue? Many Thanks!

[fgalan]

Orion 2.6.1 is pretty old. I'd suggest you to upgrade to Orion 4.2.0 and check if the problem still happening in that version.

Thanks for the feedback!