Install auto-recovery task fails with LCMVIDM74048 when attempting to patch according to KB426230

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

• When attempting to install Patch 7/8 and CSP-102547, it fails during Step # 10 when trying to run the Patch Postgres Cluster Operation with:
  • LCMVIDM74048: Unable to upload vIDM pgpool tar file to the vIDM Host
• When pinging appliances from one to another, you see a major spike in latency.
• When running arp -a you see a number of failed attempts getting to the host quickly.
• Errors in the logs appear similar to:

com.vmware.vrealize.lcm.vidm.common.exception.VidmCommandExecutionException

      at com.vmware.vrealize.lcm.vidm.clustering.util.VidmFileChangeUtil.getPgpoolDBPassword(VidmFileChangeUtil.java:414)

Caused by: java.lang.NullPointerException

Environment

VMware Aria Suite Lifecycle 8.18.0 

VMware Identity Manager 3.3.7

Cause

DNS resolution is timing out. When attempting to ping from one vIDM host to another or LCM to vIDM, a major delay is seen. This ultimately causes LCM to timeout when attempting to transfer the tarball over the network even though some attempts to connect continue to work on the patch attempts. Because this operation is network dependent, it heavily leverages DNS and any DNS outages or delays can lead to this issue.

Resolution

Prerequisites

• Backups of the /etc/hosts file or snapshots.

Procedure

1. Add each vIDM host and LCM host to each systems /etc/hosts file.
2. Reattempt the full patching procedure as outlined in https://knowledge.broadcom.com/external/article/426230.
3. Work with your networking and DNS team to isolate the issues with this core component.