Preventing new file allocations on storage target¶
The allocation of new chunk files on storage targets can be prevented by making BeeGFS assume that
there is no space left on the target. Based on this, the target will be moved to the emergency
capacity pool, where it will not be used for new files while there are other storage targets
still available in the normal or low pool. To do this, you can create a file named
free_space.override with the contained number representing the amount of free space, so
If your storage target directory (
/mnt/myraid/beegfs_storage, then you would use this:
# echo 0 > /mnt/myraid/beegfs_storage/free_space.override
This does not prevent writes or growth of existing files stored on the target.
It will take a few minutes until the information is updated. You can check the pool assignment of a
storage target by looking at the corresponding column of
beegfs-df tool output. When you want to
restore everything back to normal later, you can simply delete the
free_space.override file. All
this can be done at runtime without restarting any of the services.