Inquiry Regarding Upgrade from CHT Version 4.8.0 to 4.11

Technical Support
1

Hello Medic,

I am reaching out to seek clarification and assistance regarding an unexpected behaviour observed with our CHT application deployment. Recently, the application upgraded itself from version 4.8.0 to 4.11 without any intervention or action initiated from our end. This behaviour was entirely unanticipated.

What makes this situation particularly perplexing is that the current container images running on our Kubernetes cluster are still version 4.8.0. This raises questions about how the upgrade occurred and whether the application behavior reflects the expected operation for such a configuration.

Could you kindly shed light on the following points?

  1. Is it possible for the CHT application to upgrade automatically under any circumstances?
  2. What might explain the apparent mismatch between the application version (4.11) and the container images (4.8.0) currently in use?
  3. Are there specific configurations or settings we might review to prevent unexpected upgrades in the future?
  4. Additionally, if it is possible to identify the user or process that initiated this upgrade, please point me to where I can find such information. This would greatly aid in troubleshooting and addressing this situation effectively.

For reference, below is the description snippet of the pod in question:

Name:             cht-couchdb-1-xxxxxx-xxxxx
Namespace:        xxxxxxx
Priority:         0
Node:             ip-xxxxxx.ec2.internal/xxx.xxx.xxx
Annotations:      kubectl.kubernetes.io/restartedAt: 2024-11-06T09:58:35+03:00
Status:           Running

Controlled By:  ReplicaSet/cht-couchdb-1-xxxxxx-xxxxx
Containers:
  cht-couchdb-1:
    Container ID:   containerd://ba57ba49911c457da07ade630016d9f317c8ff5f1e179b6ae97165aac39556f2
    Image:          public.ecr.aws/medic/cht-couchdb:4.8.0
    Image ID:       public.ecr.aws/medic/cht-couchdb@sha256:9cb537b8f2f3ed1499e709a14a68981c330af105bf577a139e8129810c7ff0b2
    Port:           5984/TCP
    Host Port:      0/TCP
    State:          Running

I would greatly appreciate your guidance on this matter and any recommendations for troubleshooting or resolution. Please let me know if further details are required to assist with the investigation.

Many thanks.

2

Hi @Steve_Roy

1 & 2. The CHT doesn’t automatically upgrade under any circumstances.
3. You can check API and haproxy access logs to see when the upgrade was initiated. The upgrade could have also been initiated through K8S, at infrastructure level, by updating the helm charts that control the deployment. Since I don’t know your deployment, I have no guidelines to provide.

From your shared log, you are only showing the CouchDb pod, can you please show information about all CHT pods?
All pods matter, but the one that actually determines which code runs on the client is CHT API.

Thanks!

2 Likes