Extreme Networks ExtremeWare Version 7.8, Troubleshooting Manual

Page: 69 / 122

Per-Slot Packet Memory Scan on BlackDiamond Switches
Advanced System Diagnostics and Troubleshooting Guide 69
modules will trigger a reboot if the failed module is the master MSM. A failed MSM-64i in the slave slot
is simply removed from service.
In general, network redundancy protocols will work to recover the network. The impact on the network
depends on the network topology and configuration (for example, OSPF ECMP versus a large STP
network on a single domain).
In an OSPF network, for example, after the shutdown/reboot is initiated, the adjacent OSPF routers will
drop routes to the faltering switch. Very little traffic loss should occur during the network
reconvergence, because traffic is simply routed around the affected switch via pre-learned routes.
Because of this quick reconvergence factor, the reboot triggering a second reconvergence might be
absorbed as long as small network outages are acceptable. If no amount of loss is acceptable, the
shutdown
option should be selected, so that a maintenance window can be scheduled to reinsert the
system.
As an example of impact in an STP network, the STP network will detect the outage and trigger a
spanning tree reconvergence. The shutdown option keeps the system out of service, so no second
reconvergence will occur, keeping service outages to a minimum. This redundancy protocol is designed
to be slow and methodical to ensure that no loops exist in the network. While this protocol works very
well, administrators might wish to reinsert the system during a scheduled maintenance window. If the
reboot option is selected, once the system has recovered, a second spanning tree reconvergence will be
triggered and connectivity will be restored upon completion.