Unable to manage Edge following removal of network: Transaction was marked for rollback only; cannot commit
search cancel

Unable to manage Edge following removal of network: Transaction was marked for rollback only; cannot commit

book

Article ID: 314270

calendar_today

Updated On:

Products

VMware NSX

Issue/Introduction

  • You are using NSX for vSphere
  • A Logical Switch was deleted from NSX which was referenced in the edge firewall
  • Attempting to redeploy the edge fails with error: Transaction was marked for rollback only
  • Changing the edge logging level in the UI fails with error: Failed to publish configuration on NSX Edge appliance: null
  • Failed jobs show the below error in the UI:
  • In /logs/management-services/vsm.log you also encounter entries similar to the below:
  • Virtual wire not found in DB 

         2022-01-18 15:44:28.224 GMT ERROR TaskFrameworkExecutor-1 FwVsePojoConverter:550 - - [nsxv@6876 comp="nsx-manager"                             errorCode="MP10316" level="ERROR" subcomp="manager"] Replaced ipset virtualwire-XXX with empty ipset. Ipset virtualwire-XXX Not found           in DB

  • Sync job failed, rolling back 

         2022-01-18 15:44:28.255 GMT ERROR TaskFrameworkExecutor-1 OutOfSyncTask:88 - - [nsxv@6876 comp="nsx-manager" errorCode="MP10297"               level="ERROR" subcomp="manager"] Sync Job failed to sync edge edge-XX. Error Transaction was marked for rollback only; cannot commit;           nested exception is org.hibernate.TransactionException: Transaction was marked for rollback only; cannot commit

  • JPA Transaction error

         2022-01-18 12:36:01.947 GMT ERROR TaskFrameworkExecutor-25 OutOfSyncTask:84 - Sync Job failed to sync edge edge-1254. Error Could not           commit JPA transaction; nested exception is javax.persistence.RollbackException: Transaction marked as rollbackOnly 

 
NOTE: The preceding log excerpts are only examples. Date, time and environmental variables may vary depending on your environment.

Environment

NSX for vSphere 6.3.x

NSX for vSphere 6.4.x

Cause

This bug is caused when the logical switch is still consuming an Ipset in the NSX DB after deletion

Resolution

This issue is fixed in NSX or vSphere 6.4.6

Workaround:
To resolve this issue restart the NSX manager

To reboot the NSX manager follow any of the following steps:

  1. Log into NSX UI, click on the cog in the top right corner and press "Reboot Appliance" 
  2. SSH into the NSX manager and run the following command: reboot