[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]
Re: [atomic-devel] [demo patch] Changing the host's mount namespace from a container
- From: Daniel J Walsh <dwalsh redhat com>
- To: Colin Walters <walters verbum org>, atomic-devel projectatomic io
- Subject: Re: [atomic-devel] [demo patch] Changing the host's mount namespace from a container
- Date: Mon, 23 Feb 2015 16:13:41 -0500
On 02/23/2015 04:11 PM, Colin Walters wrote:
> I should note this patch works with "builtin" types bind mounts and direct block IO devices. But it won't help if the mount calls into an external binary, like how nfs calls into /usr/bin/mount.nfs. Each helper program would need to gain a --in-ns-of argument, and util-linux would have to pass it on. So that's a next step here.
>
I think we have to work on turning off of the --mnt namespace.
Basically just mount up the directory and run a process within the
chroot. This would eliminate all of the problems we are seeing with mnt
namespace.
[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]