As far as I understand this is not supported. What we do is we add a
new device, we then do sg_reset -b /dev/sgX and then we add them into
the volume group and then we resize it.
I would also warn you that we have seen corruption repeatable times
with xfs over devicemapper under 2.6.
Cheers
Arthur
-----
CTO @ Fotango Ltd
+447834716919
http://www.fotango.com/
On 1 Mar 2005, at 21:03, Andy wrote:
I am using dm-multipath on non-partitioned (xfs) devices on a SAN. If
I
resize the device, and rescan my path devices (i.e rescan /dev/sda,
/dev/sdb, /dev/sdc,/dev/sdd (all the same path)) the kernel sees the
new
size of the individual devices, but the /sys/block/dm-0 device is
still the
old size. I believe this is the reason why I cannot grow my xfs
filesystem
without doing a reboot so that the dm device will have the new size.
I'd
like to avoid unmounting and rebooting to resize the device. I haven't
played with lvm enough to feel comfortable using it. Is there any way
to
get the dm device to be the proper size, or is this a bug, or a
feature that
needs further discussion?
Andy
--
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel