Pivotal Knowledge Base


Ops Manager VM - /tmp Directory is Growing in Size


Pivotal Cloud Foundry® (PCF) Ops Manager versions 1.7, 1.8, 1.9, and 1.10


On the Ops Manager virtual machine (VM) the /tmp directory is growing in size.

This can be caused by frequent backups of Ops Manager, either manually by exporting the installation or by using CFOPS.


In Ops Manager 1.7 there is a daily task to clear down all files from /tmp that are not owned by the root user. This should clear down the directories created by tempest-web after an export of the OPs Manager installation.

If the service that runs this action is stopped, the /tmp directory will fill up over time. 


The first thing to check is the production logs to see if the clear down process is running. The logs can be found here /tmp/logs/production.log.

The process is called delayed_job and should appear in the logs, for example:

2016-06-21T18:06:31+0000: [Worker(delayed_job host:pivotal-ops-manager pid:1277)] Job ActiveJob::QueueAdapters::DelayedJobAdapter::JobWrapper (id=268) COMPLETED after 0.0028

If the logs show that the job is no longer running or has not run recently, then check that the tempest-scheduler is running:

service --status-all

You should see the tempest-scheduler running [+]

[ + ]  tempest-scheduler

[ + ]  tempest-web

[ + ]  tempest-worker

[ + ]  udev

If the tempest-scheduler is not running then start it by running:

service tempest-scheduler start

If the clear down process is still not running and /tmp is not being cleared, then contact Pivotal support for further investigation and advice.

Additional Information

Please see Ops Manager VM - no space left on the device for troubleshooting the issue in Ops Manager 1.6.


Powered by Zendesk