How to use iManager upgrade installation package

Author: yangjunlin

Preface

Next, I will lead you to upgrade 10.2.1imanager for k8s to version 11.0.1. (Tip: If you are upgrading the production environment, please contact the support center first to confirm the upgrade method and steps!) For details and refer to the help document
https://help.supermap.com/iManager_K8S/1101/zh/guides/UpgradePackage/

1. Prepare the environment for deploying imanager for k8s

First, let’s take a look at the imanager I installed and the cloud package I created. I use
the supermap-imanager-for-kubernetes-10.2.1-211222-548-linux-x64.tar.gz installation package.

2. Environment preparation before upgrade

1. All packages required for upgrade are at
   https://pan.baidu.com/s/1lWLyyT1jAbADGuxEG3m4gA Extraction code: ge4d
2. Download the offline image package version to be upgraded (optional). If you use the offline package, select it (after downloading the image package, copy the image package to any directory of any machine with Docker service installed, decompress it, and enter the decompressed directory. Refer to the readme.txt in the directory to synchronize the image package to the original image warehouse). If the Alibaba Cloud image is pulled online, no other operations are required.
   
   
3. Download the corresponding version of imanager for k8s upgrade package and imanager for k8s deployment package
   
4. Please read the notes in the help documentation carefully before upgrading.

2.Upgrade imanager

Note that this time I used the Alibaba Cloud image, so no offline image was installed. If it is an offline image, you also need the 11.0.1 version of the offline image. You can delete the previous 10.2.1 todel container first, and then execute the installation command. This can avoid port conflicts and container name conflicts that cause the image warehouse container to fail to start normally. Condition

1. Download the official deployment package of the version to be upgraded and unzip it
   
   
2. Enter the decompressed deployment package and modify the values.yaml file in the 11.0.1 deployment package. The contents of the values.yaml file need to be the same as the values.yaml field in the original 10.2.1 deployment package (the deploy_*_tag field should use the contents of the new deployment package). For fields that exist in the new deployment package but are not in the original deployment package, please refer to the deployment guide in the new deployment package to configure them.
   
3. Create the yamls directory in the upgrade package directory and put the values.yaml modified in 11.0.1 into it
   
4. Then enter the following 4 environment variables in command mode
   
   
5. Then start the execution.
   
   
   When the output result is the above result, it proves that the imanager upgrade is successful. Log in to the 31100 interface and re-apply for the 11.0.1 license to import.
   
6. After the upgrade is completed, please visit the iManager for k8s main page and check whether the storage management contains new and unmounted services. Unmounted services can be bound through the "bind" operation on the page.

Of course it’s not finished yet, the cloud suite site still needs to be upgraded.

3. Upgrade cloud suite

1. To ensure a smooth upgrade process before upgrading the cloud suite, please scale the StatefulSet keycloak to 1 before upgrading. If the keycloak is not a StatefulSet type, no processing is required. You can check whether keycloak is a StatefulSet by executing the following command. Execute kubectl -n get statefulset keycloak to check whether keycloak is a StatefulSet and whether the replica set is 1.
   If not, please execute kubectl -n scale --replicas=1 statefulset/keycloak
   and then use the first command to confirm whether Scale to 1
2. When the number of keycloak pods is 1, execute the following command to upgrade as last time. namespace is the namespace name of the cloud suite.
   
   
   
   
   
3. It seems to be ok, and the data has not been lost, but there are still service nodes that are down. This is because we still need to manually modify the image corresponding to the service node. No modification was made during the upgrade, and it was modified to the 11.0.1 version of the service. Node image
   
   
   and then wait for startup, everything is normal now
   

如果您是生产环境升级，请先联系支持中心，确认升级方式及步骤！