Cisco Cisco IOS Software Release 12.2(27)SBC
682
Caveats for Cisco IOS Release 12.2(28)SB through 12.2(31)SB14
OL-9967-04 Rev. A1
Caveats
•
CSCek77555
Symptoms: PPP may not start on a serial interface that is physically up. When this situation occurs,
inspection of the interface via the show interface command shows that the physical layer is up, but
that the line protocol is down, and that LCP is closed.
inspection of the interface via the show interface command shows that the physical layer is up, but
that the line protocol is down, and that LCP is closed.
Conditions: This symptom is observed only on regular serial interfaces that use PPP encapsulation.
The symptom does not occur with tunneling mechanisms such as PPP over ATM (PPPoATM) or
VPDN sessions. The symptom may occur when the physical layer undergoes multiple state
transitions, starting from an up state and ending in an up state, with the entire sequence occurring
over a short period of time. In such a situation, event filtering mechanisms in Cisco IOS software
may prevent a notification from being sent to PPP when the link returns to an up state, and, in turn,
PPP from (re-)starting on the interface. The most likely time for such a situation to occur is when
PPP itself resets the interface, which occurs when an existing PPP session is terminated because of
a keepalive failure or LCP negotiation failure.
The symptom does not occur with tunneling mechanisms such as PPP over ATM (PPPoATM) or
VPDN sessions. The symptom may occur when the physical layer undergoes multiple state
transitions, starting from an up state and ending in an up state, with the entire sequence occurring
over a short period of time. In such a situation, event filtering mechanisms in Cisco IOS software
may prevent a notification from being sent to PPP when the link returns to an up state, and, in turn,
PPP from (re-)starting on the interface. The most likely time for such a situation to occur is when
PPP itself resets the interface, which occurs when an existing PPP session is terminated because of
a keepalive failure or LCP negotiation failure.
Workaround: Any sequence that resets the physical layer and that is slow enough that the filtering
mechanisms do not once again intrude is sufficient to restart PPP. For example, you can restart PPP
on the interface by entering the shutdown interface configuration command followed by the no
shutdown interface configuration command.
mechanisms do not once again intrude is sufficient to restart PPP. For example, you can restart PPP
on the interface by entering the shutdown interface configuration command followed by the no
shutdown interface configuration command.
•
CSCek78237
Symptoms: A short CPU hog seen in the ATM PA Helper process when an interface flaps and the
framing configuration is modified on the interface.
framing configuration is modified on the interface.
Conditions: This symptom is observed on a Cisco 7200 with a PA-A3-T3 adapter that is running
Cisco IOS Release 12.2(25)S or 12.2(31)SB (and possibly other Cisco IOS releases).
Cisco IOS Release 12.2(25)S or 12.2(31)SB (and possibly other Cisco IOS releases).
Workaround: There is no workaround.
Further Problem Description: The CPU hog is enough to cause OSPF adjacencies (with fast hello)
to go down on other unrelated interfaces. The same problem is seen if BFD is configured.
to go down on other unrelated interfaces. The same problem is seen if BFD is configured.
•
CSCin99778
Symptoms: An ISG configured for RADIUS proxy may see a dummy RADIUS proxy context being
created when an accounting stop packet is received.
created when an accounting stop packet is received.
Conditions: This symptom is observed when accounting forwarding is configured and an accounting
stop comes when no subscriber session exists for that user.
stop comes when no subscriber session exists for that user.
Workaround: There is no workaround. However, this is a negative case where the ISG receives a stop
record before a start record.
record before a start record.
•
CSCir01449
Symptoms: A router that functions under a heavy load with SSHv2 clients may crash if any of the
SSH clients are terminated.
SSH clients are terminated.
Conditions: This symptom is observed on a Cisco 7600 series when the following conditions are
present:
present:
–
The CPU usage is above 70 percent.
–
There are continuous sweep pings from two far-end routers that have the debug ip packet
command enabled to create continuous logs for the SSH clients.
command enabled to create continuous logs for the SSH clients.
–
The no logging console command is configured.
–
A connection is made from a couple of SSHv2 clients, you enable the terminal monitor
command, and you terminate the SSHv2 clients while continuous messages are being generated.
command, and you terminate the SSHv2 clients while continuous messages are being generated.
–
The TCP window size is reduced.