Solved: Moving a VM into a vAPP is horrible slow when vApp contains a lot VMs

[moellerjot]

Hi vCD Fans, ✋
we have an Issue by moving a VM into a vAPP.
It takes up to 1h🕐 to complete the Recomposing, if the vAPP contains a lot VMs. 
Small vAPPs don't have this Problem. 
Does anyone have the same or a similar problem? 
Tips advice welcome! 🙂

TASK: 
ℹ️"Recomposing Virtual Application VAP_My_vApp(31d0ab24-d292-caffee-905a-73a3affebeef) " Resuming BlockingTaskℹ️

Actions Done: 
- vCD RAM + CPU + Java Heap space resizing -> done
- Database Cleanup -> done 
- GCS Ticket Prio3 -> open 

Environment: 
vCD 10.2 + MQTT + vRO8.3 (TCI 2.0)

My Suspicion: 
Recursive database calls that increase exponentially due to the number of VMs in a vAPP.

[moellerjot · Accepted solution]

The Issue was cased by a vCD sql statement that requests a portgroup in vCenter that do not exist anymore.  

tail -f /opt/vmware/vcloud-director/logs/vcloud-container-debug.log | grep "Unable to"

2021-12-10 09:50:18,315 | DEBUG    | Backend-activity-pool-191932 | NetworkBackingToSwitchMapper   | Unable to find dvs for portgroup ....