https://bugzilla.novell.com/show_bug.cgi?id=250620 Summary: ifdown may freeze with bonding driver - patch provided Product: openSUSE 10.2 Version: Final Platform: i386 OS/Version: SuSE Other Status: NEW Severity: Normal Priority: P5 - None Component: Basesystem AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: eugen@drnet.at QAContact: qa@suse.de ifdown as distributed may fail on bonded interfaces: We have seen "ip link set dev $INTERFACE down &>/dev/null" in line 1201 never return after bonding a r8169 primary with a compaq NC3134 dualport e100 secondary. This prevents a clean shutdown. AFAI have tested the problem appears only if - there are other interfaces besides the bond - the bond is the only interface carrying a default route - there is at least one e100 and at least one other card in the bond This might be a kernel issue more than a distribution issue. When calling "ifenslave -d" on all slaves before the "ifdown" this problem goes away. the following patch calls "ifenslave -d" on all slaves before downing the bonding instance, this solved the problem for our setup. Hope the web form doesnt ruin linebreaks and whitespace - mail me up if this is the case. fwright:/sbin # diff -u ifup ifup.bonding-patched --- ifup 2006-11-25 13:55:27.000000000 +0100 +++ ifup.bonding-patched 2007-03-02 15:16:35.000000000 +0100 @@ -1195,6 +1195,13 @@ ifup-route $CONFIG $INTERFACE ${OPTIONS:+-o $OPTIONS} ;; ifdown) + if [ "$BONDING_MASTER" = yes ] ; then + SLAVES=`cat /proc/net/bonding/$INTERFACE | grep "Slave Interface" | awk '{print $3;}'` + for SLAVE in $SLAVES; do + message " Unenslaving interface $SLAVE" + ifenslave -d $INTERFACE $SLAVE &>/dev/null + done + fi ifdown-dhcp $CONFIG $INTERFACE ${OPTIONS:+-o $OPTIONS} ifdown-route $CONFIG $INTERFACE ${OPTIONS:+-o $OPTIONS} ip addr flush dev $INTERFACE &>/dev/null -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.