11.4. Recovering from a Full Disk
The following procedure describes how to resolve the situation when a logical volume (LV) with the Pulp database on it has no free space.
To recover from a full disk
- Let running Pulp tasks finish but do not trigger any new ones as they can fail due to the full disk.
Ensure that the LV with the
/var/lib/pulpdirectory on it has sufficient free space. Here are some ways to achieve that:Remove orphaned content:
# foreman-rake katello:delete_orphaned_content RAILS_ENV=production
This is run weekly so it will not free much space.
- Change the download policy from Immediate to On Demand for as many repositories as possible and remove already downloaded packages. See the Red Hat Knowledgebase solution How to change syncing policy for Repositories on Satellite from "Immediate" to "On-Demand" on the Red Hat Customer Portal for instructions.
Grow the file system on the LV with the
/var/lib/pulpdirectory on it. For more information, see Growing a File System on a Logical Volume in the Red Hat Enterprise Linux 7 Logical Volume Manager Administration Guide.注記If you use an untypical file system (other than for example ext3, ext4, or xfs), you might need to unmount the file system so that it is not in use. In that case, complete the following steps:
Stop the
satellite-maintainservices:# satellite-maintain service stop
- Grow the file system on the LV.
Start the
satellite-maintainservices:# satellite-maintain service start
- If some Pulp tasks failed due to the full disk, run them again.