Hello,
I've been exploring FreeBSD for a few weeks. I am trying to do something that can be done easily in Linux docker or Solaris/illumos zone, but failed to find equivalent way of doing so in FreeBSD jail.
Suppose in the host I have the following mountpoints:
/vmgr
/vmgr/subdir1
/vmgr/subdir2
/vmgr/subdir3
...
All of the above are different zfs file systems (or datasets). New file systems could be created and old ones removed under /vmgr.
All I want to do is to have the jail instances see what the host can see, i.e. all under /vmgr, across file systems. zfs datasets delegation is not an option, because I want the host, and all jail instances all see the same mount points/directory structure.
In Linux, this can be done by specifying '--mount bind-propagation=shared,type=bind,source=/vmgr,target=/vmgr' when starting the docker instance; in Solaris/illumos, this can be achieved by mounting /vmgr through lofs (loopback file system) in the zone.
I read nullfs manual, it does not seem to help, because it won't cross file system boundaries.
Is this possible in FreeBSD?
Thanks,
--Youzhong
I've been exploring FreeBSD for a few weeks. I am trying to do something that can be done easily in Linux docker or Solaris/illumos zone, but failed to find equivalent way of doing so in FreeBSD jail.
Suppose in the host I have the following mountpoints:
/vmgr
/vmgr/subdir1
/vmgr/subdir2
/vmgr/subdir3
...
All of the above are different zfs file systems (or datasets). New file systems could be created and old ones removed under /vmgr.
All I want to do is to have the jail instances see what the host can see, i.e. all under /vmgr, across file systems. zfs datasets delegation is not an option, because I want the host, and all jail instances all see the same mount points/directory structure.
In Linux, this can be done by specifying '--mount bind-propagation=shared,type=bind,source=/vmgr,target=/vmgr' when starting the docker instance; in Solaris/illumos, this can be achieved by mounting /vmgr through lofs (loopback file system) in the zone.
I read nullfs manual, it does not seem to help, because it won't cross file system boundaries.
Is this possible in FreeBSD?
Thanks,
--Youzhong