We have a small issue on the new trading floor switches; sclisgh011, sclisgh012, sclisgh111, and sclisgh112 (all Cat 6500, dual-redundant supe 720, with10G uplinks and 240 * 1G access ports).
These devices are configured as layer-2 switches, and include the command "no ip routing". Each switch is managed by telnet via an IP addresses configured on management vlan interface (and a ip default gateway command).
Whenever the devices are power cycled they lose their "no ip routing" configuration and start routing, and so management connectivity is lost (because the devices don't have an ospf configuration).
To overcome this issue a static default route has been added to each box, pointing to the default gateway.
The same issue (ip routing) arises when ever the boxes are failed-over to the hot-standby redundant supervisor. The boxes run SSO with NSF and traffic forwarding is not affected, only management of the boxes.
Any idea why this command is disapearring ? Is this a bug ?