Patching of the vCenter Server is stuck, and the RPM installation does not complete.

Products

• VMware vCenter Server

Issue/Introduction

• The update process seems stuck on RPM installation with below messages:
  
  update_microservice.log
  update_b2b_target::            executeInstallRpms:1938 -     INFO - Installation of Photon RPMS completed successfully.
update_b2b_target::                  _isDowngrade:2254 -     INFO - It is downgrade: False
functions_target::               runInstallVCRpm: 166 -     INFO - Installing VC RPMS.functions_target::                    runCommand:  82 -    DEBUG - Invoking command [/bin/bash --login -c /bin/rpm  -Uhv /storage/updatemgr/software-


• The installation does not appear to have started.
  
  message.log
VC [RPM][22724]: erase VMware-spherelet-2.1.2-13442867.x86_64: success
VC [RPM][22724]: erase vmware-snmp-1.8.0-1.x86_64: success
VC [RPM][22724]: erase vmware-sca-8.0.1.00500-13442867.noarch: success

Environment

• Patching vCenter Server 8.x　(8.0u1e to 8.0u2e)

Cause

• The RPM upgrade hung in the following sleep loop within the lwis upgrade script if stale lwsmd processes were running simultaneously and lwsm status reports an erros.
  
  ------------------------------------------------------------------------------------------------
  root         1  9906  0.0  0.0 518000  4548 ?       Sl    2024 00:01:25     35    642 /opt/likewise/sbin/lwsmd --start-as-daemon --syslog
root         1 17233  0.0  0.0 518000  4856 ?      Sl   Jun19 00:00:48     13    610 /opt/likewise/sbin/lwsmd --start-as-daemon --syslog
root         1 46178  0.0  0.0 591732  4496 ?      Sl    2024 00:00:50     55    776 /opt/likewise/sbin/lwsmd --start-as-daemon --syslog
  
  
  update_microservice.log.1
  functions_target:: runCommand:82 - DEBUG - Invoking command [/bin/bash --login -c /bin/rpm  -Uhv /storage/updatemgr/software-updates25jajx7/stage/*.rpm]
  
  ps_auwwx---sort--rss.txt
  root 22724 0.0 0.0 40060 23400 pts/1 S+ 16:53 0:01 /bin/rpm -Uhv /storage/updatemgr/software-updates25jajx7/stage/applmgmt-8.0.3.00600-15529252.x86_64.rpm /storage/...



Resolution

The engineer is aware of the issue, and a timeout will be implemented in a future release if it cannot be started.

As workaround, restore vCenter Server from snapshot or backup and verify that only a single lwsmd process is running before retrying the vCenter Server upgrade. (In this example, 1596 is the PID of the single running process. Your output may differ.)

pidof lwsmd
1596