[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#409042: dmsetup executed even if not available



On Tue, Jan 30, 2007 at 01:29:52PM +0100, Frans Pop wrote:
While looking through the syslog with this installation report, I noticed
the following lines (repeated) in the syslog:
Jan 24 15:58:11 main-menu[2105]: (process:13045): /lib/partman/choose_method/50lvm/choices: 20: Jan 24 15:58:11 main-menu[2105]: (process:13045): dmsetup: not found
David, could you have a look at this?
It would be nice if this could be fixed before Etch.

Seems to originate in partman-lvm/choose_method/lvm/choices.

Yes. Grepping through partman-lvm gives only one invocation of dmsetup. The following lines:

elif grep -q "/dev/mapper/" $dev/device; then
	# LVM on device-mapper crypto
	device=$(cat $dev/device)
	if [ "$(dmsetup status $device | cut -d' ' -f3)" = "crypt" ]; then
		lvm=yes
	fi
fi

The problem is that dmsetup-udeb is a dependency of cryptsetup-udeb which is loaded on demand by partman-crypto.

But, since the test is for whether a crypto device has been setup, a missing dmsetup also means that there is no cryptsetup present and that no crypto device has been setup. Therefore, it will be enough to guard the dmsetup invocation with a test that dmsetup also exists.

So I've committed the following fix to partman-lvm:

Index: choose_method/lvm/choices
===================================================================
--- choose_method/lvm/choices   (revision 44677)
+++ choose_method/lvm/choices   (working copy)
@@ -13,9 +13,11 @@
       lvm=yes
elif grep -q "/dev/mapper/" $dev/device; then
       # LVM on device-mapper crypto
-       device=$(cat $dev/device)
-       if [ "$(dmsetup status $device | cut -d' ' -f3)" = "crypt" ]; then
-               lvm=yes
+       if type dmsetup > /dev/null 2>&1 ; then
+               device=$(cat $dev/device)
+               if [ "$(dmsetup status $device | cut -d' ' -f3)" = "crypt" ]; then
+                       lvm=yes
+               fi
       fi
fi


--
David Härdeman



Reply to: