CentOS - Dec 2016 - Can't delete or move /home on 7.3 install

On Thu, Dec 15, 2016 at 11:51:28AM -0800, Glenn E. Bailey III
wrote:
> Doing a lsof showed no open files against /home. Something else is
> locking it, not a user process. Also disabled SELinux, did a init 1,
> and only way to remove it was via single user by passing
> init=/sysinit/bin/sh
It sounds like /home is being managed by something in the kernel, then.

Are you exporting /home via NFS or SMB?  AutoFS?  I'm trying to think
what would start at runlevel 1, although if this is c7 that might not
be a valid description anymore, perhaps services are being started at
'rescue.target'.

-- 
Jonathan Billings <billings at negate.org>

This is a base install. If you deploy an instance in ec2 or GCE (ec2
you can do the free tier) it's easily repeatable. Even on a RHEL 7.3
instance. Note you'll need to allow root and password logins via SSH
before attempting.

On Fri, Dec 16, 2016 at 5:47 AM, Jonathan Billings <billings at
negate.org> wrote:
> On Thu, Dec 15, 2016 at 11:51:28AM -0800, Glenn E. Bailey III wrote:
>> Doing a lsof showed no open files against /home. Something else is
>> locking it, not a user process. Also disabled SELinux, did a init 1,
>> and only way to remove it was via single user by passing
>> init=/sysinit/bin/sh
>
> It sounds like /home is being managed by something in the kernel, then.
>
> Are you exporting /home via NFS or SMB?  AutoFS?  I'm trying to think
> what would start at runlevel 1, although if this is c7 that might not
> be a valid description anymore, perhaps services are being started at
> 'rescue.target'.
>
> --
> Jonathan Billings <billings at negate.org>
> _______________________________________________
> CentOS mailing list
> CentOS at centos.org
> https://lists.centos.org/mailman/listinfo/centos


-- 
"replicants are like any other machine. They're either a benefit or a
hazard. If they're a benefit, it's not my problem."

On Fri, Dec 16, 2016 at 10:17:21AM -0800, Glenn E. Bailey III
wrote:
> This is a base install. If you deploy an instance in ec2 or GCE (ec2
> you can do the free tier) it's easily repeatable. Even on a RHEL 7.3
> instance. Note you'll need to allow root and password logins via SSH
> before attempting.
I can confirm this.

The culprit?  NetworkManager has /home open.  I can't figure out
*WHY*.

# systemctl start NetworkManager.service
# rmdir /home
rmdir: failed to remove ?/home?: Device or resource busy
# systemctl stop NetworkManager.service
# rmdir /home
rmdir: failed to remove ?/home?: Device or resource busy
# systemctl stop NetworkManager.service
# rmdir /home
rmdir: failed to remove ?/home?: Device or resource busy
# systemctl kill NetworkManager.service
# rmdir /home
rmdir: failed to remove ?/home?: Device or resource busy
# rmdir /home

So it takes some time for all the parts of NetworkManager to die, but
eventually, they do and you can remove /home.

Unfortunately, I don't see any reason why /home would be in use by
NM.   Very odd.

-- 
Jonathan Billings <billings at negate.org>