iSCSI and root

iSCSI and root

[Harald Hoyer]

Because root=dhcp with root_path=iscsi:.. or root=iscsi:... would just add a 
harddisk to the system, we still might want a root=/dev/... to specify the path 
to the root partition.

Or should we just mount the whole disk? If so, how do we detect the newly added 
disk after iscsistart? Do we add a udev rule to mount it?

We might want to use the code from /lib/udev/path_id, which parses 
$full_sysfs_device_path and scans for  */session[0-9]*/*

Any suggestions?
--
To unsubscribe from this list: send the line "unsubscribe initramfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: iSCSI and root

[Harald Hoyer]

On 06/02/2009 06:10 PM, Harald Hoyer wrote:
> Because root=dhcp with root_path=iscsi:.. or root=iscsi:... would just
> add a harddisk to the system, we still might want a root=/dev/... to
> specify the path to the root partition.
>
> Or should we just mount the whole disk? If so, how do we detect the
> newly added disk after iscsistart? Do we add a udev rule to mount it?
>
> We might want to use the code from /lib/udev/path_id, which parses
> $full_sysfs_device_path and scans for */session[0-9]*/*

or just try to mount everything /dev/disk/by-path/ip-*-lun-*

>
> Any suggestions?
> --
> To unsubscribe from this list: send the line "unsubscribe initramfs" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe initramfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: iSCSI and root

[David Dillow]

On Tue, 2009-06-02 at 18:15 +0200, Harald Hoyer wrote:
> On 06/02/2009 06:10 PM, Harald Hoyer wrote:
> > Because root=dhcp with root_path=iscsi:.. or root=iscsi:... would just
> > add a harddisk to the system, we still might want a root=/dev/... to
> > specify the path to the root partition.
> >
> > Or should we just mount the whole disk? If so, how do we detect the
> > newly added disk after iscsistart? Do we add a udev rule to mount it?
> >
> > We might want to use the code from /lib/udev/path_id, which parses
> > $full_sysfs_device_path and scans for */session[0-9]*/*
> 
> or just try to mount everything /dev/disk/by-path/ip-*-lun-*

I was thinking that for iscsi and nbd, we could allow a root=/dev/sda or
root=LABEL=/something along with a netroot=iscsi.... or netroot=nbd to
pull in the block devices, and then allow udev to mount them as they get
added. Ie, allow re-use of the existing block rules.

This would allow people to also do LVM and encrypted PVs on the
networked block devices.

We'd keep the root=iscsi*/nbd* stuff and default to either /dev/nbd0 or
the LUN specified for iscsi.
--
To unsubscribe from this list: send the line "unsubscribe initramfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html