Cisco MGX 8850 PXM45 Release Notes page 30

Limitations, Restrictions, and Notes
AXSM Cards
•
•
•
•
•
•
•
PNNI Limitation
•
Release Notes for Cisco MGX 8850 (PXM45), MGX 8850 (PXM1E), and MGX 8830, Software Release 3.0.10
30
Do not enable offline diagnostics on AXSM cards in the APS 1+1 Op A mode. A possible loss of
signal will occur when offline diagnostics are run on the AXSM cards in this configuration (refer to
caveat CSCdy62950).
For AXSM OC-3 cards running in the Op B mode, a remote card switchover can occur by doing
setrev on a local node. This problem occurs when more than twelve lines are configured for APS
and the setrev command is issued. The mode of the AXSM card can be found by issuing the dspcd
command. The card switchover on the remote node is happening because of flooding of several
interrupts occurring on all the lines. The cause of the interrupt was the reset of both the cards on
local node (refer to caveat CSCdy09027).
APS is supported on AXSM-1-2488/B after upgrading the card to Release 3.0.00 and up, and
enabling the card to operate in the Op B mode.
The dspapsln/dspapslns/dspalms/dspalm commands on a standby AXSM card do not show the
results consistent with an active card. When the cards are operating in the Op B mode (which can
be found by issuing the dspcd command on an AXSM card), the standby card does not perform the
alarm integration and APS is also not enabled on the standby card. All the defects or failures and
statistics are dynamic events which are not sent to the standby card—the standby card is unable to
show the correct faults as shown by the active card. These commands are not blocked on the standby
card as these are useful for debugging purposes (refer to caveat CSCdx63098).
When there is an APS line between an AXSM card in the Op A mode and an AXSM card in the Op B
mode, the working and protected lines can go into the SD (signal degraded) mode if the cards are
set up for 1:1 intracard APS and SD is set to –7 upward. The problem is seen because of the behavior
of the AXSM card in the Op A mode, and the problem exists only between an AXSM card in the
Op B mode and an AXSM card in the Op A mode; it does not exist if both AXSM cards are in the
Op B mode or in the Op A mode.
Whenever a switch is made on the AXSM card in the Op A mode, it does the bridging to the other
line on which a switch is going to happen. In this duration, the AXSM card in the Op B mode notices
the change in the BER count and although the BER count is very small, it is sufficient to declare
Signal Degraded (SD) because of having the SD threshold set at –7 (refer to caveat CSCdy31269).
A standby AXSM card cannot sync up with an active AXSM card after loadrev to Release 3.0.10 if
the redundant AXSM card has NNI ports and 40k connections configured, and both cards the
2.0(15.2) image. In the 2.0.x release, the LCN could be 0 but in later releases the ATLAS LCN 0 has
been taken out from the pool. So for the cards with 4 ATLASs, the problem will be seen for the
following LCNs while upgrading to Release 3.0.10 (refer to caveat CSCdy36202):
0
–
32 * 1024 = 32768
–
64 * 1024 = 65536
–
96 * 1024 = 98304
–
There is a limitation in the ATM Forum PNNI specification on how the crankbacks are handled by
the entry border nodes. If the entry border of a peer group cannot route a call to the destination node
and if the cause of blocking was within the peer group, then the entry cranks back to the next higher
level (page 246, point b.1.2 in the ATMF PNNI specification). This higher level crankback is
translated to a blocked node of the logical group node and so the source node processing this
OL-2914-01 Rev D0, July 3, 2003