Instance Resize/Migration Failure due to low Disk Space

Problem

The resize or migration of an instance fails with the following error observed in the ostackhost log located at /var/log/pf9/ostackhost.log

TRACE nova.compute.manager [instance: UUID] Command: scp -r /opt/pf9/data/instances/e716c3f6-5433-45af-bed6-a688ad559268_resize/disk 10.32.23.242:/opt/pf9/data/instances/e716c3f6-5433-45af-bed6-a688ad559268/diskTRACE nova.compute.manager [instance: UUID] Exit code: 1TRACE nova.compute.manager [instance: UUID] Stdout: u''TRACE nova.compute.manager [instance: UUID] Stderr: u'scp: /opt/pf9/data/instances/e716c3f6-5433-45af-bed6-a688ad559268/disk: No space left on device\n'

Environment

• Platform9 Managed OpenStack - All versions
• Nova
• NFS

Cause

The resize or cold migration failure occurs due to low disk space on the mount or due to the instances mount missing on the host.

Resolution

1. Check the details for the instances mount on the source and destination compute node.

$ df -h | grep instances/dev/mapper/centos-instances 558G 556G 2G 99% /opt/pf9/data/instances

2. If the space on the NFS mount is low, add more disk resources to enable resize/migration option.
3. If the instances mount cannot be found after running the command from Step 1, you will need to mount the partition using the following command.

$ sudo mount /dev/mapper/centos-instances /opt/pf9/data/instance

Note: To persist this change, an entry needs to be added to /etc/fstab file.