Metrics registrar job is failing after upgrading TAS

Products

• VMware Tanzu Application Service for VMs

Issue/Introduction

While upgrading Tanzu Application Service tile, an error occurs updating metric_registrar_orchestror on clock_global instance:

Task 106164 | 17:55:51 | Updating instance clock_global: clock_global/23524770-79ff-48b1-80df-ade52aac7a76 (1) (canary) (00:05:45)
                      L Error: 'clock_global/23524770-79ff-48b1-80df-ade52aac7a76 (1)' is not running after update. Review logs for failed jobs: metric_registrar_orchestrator

The root cause of the issue is that the metric registrar doesn't fail gracefully when CC is unreachable: On startup, the metric registrar will fail outright when CC is unreachable.

Environment

Product Version: 2.10

Resolution

For now the next release of TAS contains the revert of this metric-registrar bump until we can address the rest of the issues. metric-registrar 1.2.1 is in TAS versions: 2.7.24, 2.8.18, 2.9.12, 2.10.4 and was reverted in TAS: 2.7.25, 2.8.19, 2.9.13, 2.10.5 

The solution is to download and upgrade to the latest build release from VMWare Tanzu Network.

The affected versions of metric-registrar is 1.2.1, which is bundled in TAS versions: 2.7.24, 2.8.18, 2.9.12, 2.10.4. The temporary fix is to revert metric-registrar in TAS: 2.7.25, 2.8.19, 2.9.13, 2.10.5. In order to resolve this issue, please upgrade TAS tile to higher build version.

The root cause of the issue is that metric registrar doesn't fail gracefully when CC is unreachable. This will cause the metric registrar orchestrator to fail outright when the CC is unreachable. 

The TAS release notes lists the bug.