Solved ZFS can't boot after dumb entry in /etc/fstab

I'm posting this just in case anyone else runs into similar troubles, or I do again and need a reference NOT on my server's wiki.

at some point I created an /etc/fstab entry like:

Code:
none                              /ram                           tmpfs      rw,size=2147483648    0       0

I'm not sure why it was "none" rather than "tmpfs" but:


Code:
Starting file system checks:
Mounting local filesystems:mount: none: Operation not supported by device
Mounting /etc/fstab filesystems failed, will retry after root mount hold release
mount: none: Operation not supported by device

Mounting /etc/fstab filesystems failed, startup aborted
ERROR: ABORTING BOOT (sending SIGTERM to parent)!

dropping to shell and commenting out the offending line permitted successful reboot.
 
This is strange because I have in my fstab(5):
none /tmp tmpfs rw,size=8G,nosuid,noexec,mode=1777,gid=0,uid=0 0 0
And it's ok with RAM as the 1st token, too. Check: mount -t tmpfs -o size=1G RAM /mnt succeeds.
IMHO the 1st token is irrelevant for tmpfs(5) (bug: should be man section 4). The problem must be somewhere else. I had RAM as the 1st token on FreeBSD 10-12.1, maybe 9, too. Did you patch any of the related rc(8) scripts?
 
none is only valid for swap.
Yes for the 2nd (mountpoint) token. But it works as the 1st (device) token for tmpfs(5), as you can easily verify.
EDIT It's ok at least for procfs(5) & fdescfs(5), too.
No, it's a filesystem. It's a virtual filesystem but a filesystem nonetheless.
intro(5): "This section contains information about file formats", most (all?) other man pages in section 5 are about file formats, i.e. mostly about the format of *.conf files. Section 4 is described by intro(4) for devices, but is commonly used for any kernel module, e.g. nfsv4(4) is in section 4. This is not so important, what counts is the quality of the man pages, and this is very good in many cases.
 
Back
Top