Skip to content

OSPool Integration Maintenance

In general, once your cluster is contributing to the OSPool, things should mostly just continue to work. There are no required, routine maintenance tasks specific to maintaining the integration.

Monitoring

If you want to monitor aspects of the integration, here are a few ideas:

  • Verify that SSH to the login node continues to work.

  • Verify that the batch system on the login node is working; for example, can our user run basic query and submit commands?

  • Check whether any of our jobs run, especially when you would expect them to.

  • If your integration is via a Hosted CE (Option 1a), check your OSPool Hosted CE Dashboard to see if the OSPool is actually getting your contributions and allocating capacity to researchers.

  • If you told us to configure a Squid proxy service, perform a periodic functional test on it: Can it actually fetch a page from offsite? Squid has a strange failure mode in which the service is running but will not respond to any attempt to actually use it.

When To Contact Us

If you have or are concerned about a possible security incident that affects your cluster or may affect our integration, please email our cybersecurity team.

For all other topics, please email our support system; a real human will reply ASAP!

Some reasons to contact us:

  • You have a scheduled or unplanned downtime. Please see here for instructions on registering that downtime with us; we will stop trying to submit jobs into your cluster until it is over (or you tell us explicitly to start again).

  • Your site changes in a way that could affect the integration (see the Requirements page); e.g., networking, login node, SSH, cluster configuration, OS distro or major version.

  • You want to change the configuration (“shape” or maximum number) of jobs that we submit into your cluster.