rancher: UI is stuck on saving "custom" cluster info in 2.4.3 - Note: Clusters can still be upgraded with docker command

After upgrading from rancher 2.3.6 to 2.4.3 kubernetes upgrade is stalling and not completing. Screen Shot 2020-05-08 at 10 43 53 AM


Useful Info
Versions Rancher v2.4.3 UI: v2.4.22
Route undefined

About this issue

  • Original URL
  • State: closed
  • Created 4 years ago
  • Comments: 27 (11 by maintainers)

Commits related to this issue

Most upvoted comments

Reproduced reliably with the following steps:

  1. Create a custom data.json for rke-metadata-config that does not have the CIS Config Params or CIS Benchmark Version Info
  2. Install Rancher with the environment variable for RKE metadata config pointing to that data.json: sudo docker run -d --restart=unless-stopped -p 80:80 -p 443:443 -e "CATTLE_RKE_METADATA_CONFIG={\"refresh-interval-minutes\":\"144000\",\"url\":\"https://raw.githubusercontent.com/rancher-max/rke-data/master/data.json\"}" rancher/rancher:master-head
  3. Attempt to create a cluster. It will fail with all the errors mentioned above:
    • Null selection in Advanced Options > Scheduled CIS Scan Enabled
    • After clicking “Create”, it just spins for ever with “Saving”
    • Same error in console as mentioned above

Validated in both master-head and v2.4-head using UI index latest2 and latest-2.4 respectively

  • Default value is set to No for Scheduled CIS Scan Enabled image

  • Creating cluster does not stall. Creates successfully

  • Upgrading cluster is successful

  • No error occurs in console logs

NOTE: This fixes cluster creation/upgrade, but the error of having no CIS Scan Profiles is still present. #27005 will address that.