Pivotal Cloud Foundry® Ops Manager versions 1.2.x, 1.3.x, and 1.4.0
When export installed products vis Ops Manager, sometimes the export process takes more than 10 minutes and finally fails with error:
500 An error occurred.
Contact Pivotal Technical Support report the problem.
When running the export, backend process on Ops Manager archives files such as stemcells, compiled packages, configurations, release packages into one .zip file. Some of those files are very large so that it takes longer to archive them. Especially on a virtual machine (Ops Manager VM) which has very limited RAM, it will take more time because of memory paging.
The current timeout setting is configured as 600s. However, in case of more than 2 or 3 services installed, it is not sufficient for export process to complete within 600s.
There are a few improvements on product side from Ops Manager v1.4.x:
- By sharing stemcells between releases, carrying them in Ops Manager instead, export archive file size is shrunk.
- The timeout of the export(and import) operation is changed from 600s to 1800s in Ops Manager v1.4.1.
As a temporary workaround, if you are using Ops Manager less than v1.4.1, please manually configure timeout value or assign more resource to Ops Manager VM to improve performance.
Configure bigger timeout value:
- SSH into Ops Manager VM with 'tempest' account.
- Stop Ops Manager process by issuing 'sudo service tempest-web stop'.
- Edit /home/tempest-web/tempest/web/config/thin.production.yml, change 'timeout' from 600 to 1800, save and quit.
- Start Ops Manager process by issuing 'sudo service tempest-web start'.
- Log in Ops Manager web console, re-run "Export Installation Settings".
Assign more system resource to Ops Manager which helps to improve performance significantly:
- Increase numbers of virtual CPU.
- Assign 4GB or 8GB RAM to Ops Manager VM if possible.