Upgrading TAS fails with "Failed Jobs: rep."

Products

• VMware Tanzu Application Service for VMs

Issue/Introduction

Installing TAS with NSX-T can fail with following error.

 Updating instance diego_cell: diego_cell/0c0b23f3-e562-42de-b0f4-cf01f562878f (21) (00:12:06)
                       L Error: Action Failed get_task: Task db4ae2fe-dba4-4412-5471-531329a2c1a1 result: 1 of 6 post-start scripts failed. Failed Jobs: rep. Successful Jobs: vxlan-policy-agent, silk-daemon, route_emitter, bosh-dns, garden.
Task 392416 | 22:55:00 | Updating instance diego_cell: diego_cell/2161c56d-dc05-4b2f-aa9b-e10c1c5b05aa (9) (00:12:07)
                       L Error: Action Failed get_task: Task 579af9f1-a3d4-46c6-7466-fdca0e0185be result: 1 of 6 post-start scripts failed. Failed Jobs: rep. Successful Jobs: silk-daemon, vxlan-policy-agent, route_emitter, bosh-dns, garden.
Task 392416 | 22:55:00 | Error: Action Failed get_task: Task db4ae2fe-dba4-4412-5471-531329a2c1a1 result: 1 of 6 post-start scripts failed. Failed Jobs: rep. Successful Jobs: vxlan-policy-agent, silk-daemon, route_emitter, bosh-dns, garden.

 

Pushing apps will fail with following error in app logs
 

2023-07-14T22:23:08.83-0400 [STG/0] ERR Cell d98638a7-7307-4409-bbd0-80f5d9fd36ea failed to create container for instance f96969a7-50c2-4bef-ae1c-fd5054d34ce7: external networker up: exit status 1

And when you ssh in diego_database VM, "monit summary" doesn't show the "ncp" process.

Environment

Product Version: 2.13

Resolution

This happens when upgrading TAS to a version that is incompatible with current NSX-T version which actually deletes NSX Container Plugin.

Installing NSX-T to a compatible version should solve the issue. 

In order to prevent the issue from happening in the future, check TAS/NSX Container Plugin Interoperability Matrix or NSX Container Plugin Release Notes -> Compatibility Requirements. E.g This is for 4.0.1