Troubleshooting storage

Causes and remedies for insufficient space, an image in use, an upload that is too large, an inaccessible datastore and refused file operations.

Every refusal comes with a message id such as error.datastore.insufficient_space. The error codes page leads from the id to the full entry.

Insufficient space

Cause. A new or grown disk asks for more bytes than the datastore has free, or a thin disk is larger than the limit that the administrator set: error.datastore.insufficient_space.

Remedy. Choose a smaller disk or a datastore with more free space. Free space by deleting files that nothing uses: kyvernon datastore orphans <datastore> lists the disk images that no virtual machine, template or snapshot references, see Disks. An administrator can raise max_thin_disk_gib for thin disks.

Image in use

Cause. A disk was to be made from an image file that is still the disk of another virtual machine that runs or is suspended: error.image_in_use. Deleting or overwriting a file that holds the disk of a virtual machine, or of one of its snapshots, is refused as well: error.datastore_file_in_use.

Remedy. Power off the virtual machine that holds the file, or delete it when it is no longer needed. Make new disks from an image that no machine uses, from a template, or clone the virtual machine instead of its file. A linked clone needs its base image to stay: delete the clones before the image.

Upload too large

Cause. The file is larger than the upload limit of the controller, max_upload_bytes in the [files] section of its configuration: error.datastore_upload_too_large. The answer has status 413.

Remedy. An administrator raises the limit and restarts the controller, or puts the file on the datastore another way, for example by copying it on the host. An upload that ended early reports error.datastore_upload_incomplete: upload the file again with kyvernon datastore upload <datastore> <file> --path <path> --overwrite.

Datastore not accessible

Cause. No connected host mounts the datastore, or the storage connection of the hosts is down: error.datastore_inaccessible and error.datastore_no_host. The console shows Datastore not accessible.

Remedy. Check the mount, the network path and the state of the storage server. Reconnect a host that was disconnected on purpose with kyvernon host reconnect <host>.

Write refused

Cause. The datastore is mounted read-only: error.datastore_read_only.

Remedy. Choose another datastore, or ask an administrator to mount it read-write.

Path refused

Cause. The path is empty, absolute, contains .., or leaves the root of the datastore: error.datastore_bad_path. A file that exists already is reported as error.datastore_file_exists, and a folder that is not empty as error.datastore_folder_not_empty.

Remedy. Use a path relative to the root, such as images/disk1.qcow2. Allow overwriting on purpose with --overwrite, or delete the contents of the folder first.