usually creating RAM disks works with the following commands
hdid -nomount ram://<blocksize>
Returns e.g. /dev/disk2 Then I would format the disk, with say
newfs_hfs /dev/disk2
followed by mounting it:
mount -t hfs /dev/disk2 /some/mount/target
This procedure doesn't seem to work with APFS. I'm on High Sierra beta 9. The mount command doesn't output any error, but the path is not mounted.
In my case, after the hdid
command finished, newfs_apfs -i /dev/disk2
yields
nx_kernel_mount:1364: checkpoint search: largest xid 1, best xid 1 @ 1
nx_kernel_mount:1422: sanity checking all nx state... please be patient.
spaceman_metazone_init:278: no metazone for device 0, of size 209715200 bytes, block_size 4096
apfs_newfs:18075: FS will NOT be encrypted.
When I then enter mount -t apfs /dev/disk2 /some/target/path
then the mount commands seems to work for 2 seconds, doesn't give any output and the mount was NOT succesful.
Can anyone tell me how to actually make a APFS RAM disk s.t. it works? :p
PS: I've also tried something like diskutil partitionDisk /dev/disk2 GPT APFS myvolumename 0b
which does mount the volume to /Volumes/myvolumename
but creates yet another disk (disk3
in this case) which seems odd to me!
Found a solution:
where
<format>
is taken fromdiskutil listFilesystems
from the "Personality" column. Yes, it seems weird to me too that you may have to quote this parameter, e.g. when specifying case-sensitive variants, but oh well...<blocksize>
is 2048 *desired size in megabytes
The last command formats the RAM disk and mounts it to
/Volumes/<diskname>
It seems to be the case that when now entering
diskutil list
that you will see two new disks, the onehdid
created, and a synthesized one.To destroy the RAM disk again, call
diskutil eject <output path of previous hdid command>
, e.g.diskutil eject /dev/disk2
This will do all the work for you, unmounting the
/Volumes/<diskname>
path and destroy the two disks, releasing your memory.Keep in mind that the minimum/maximum values for
<blocksize>
depend on the chosen<format>
. Also,<diskname>
cannot always be chosen arbitrarily. Exemplary, FAT32 requires it to consist of upper-case letters!Cheers!