mail.net.cisco.bba at spam
Aug 30, 2005, 2:28 PM
Post #1 of 1
(840 views)
Permalink
|
|
7200 PPPoA terminator limits the amount of sessions open
|
|
Hi,
Upgrading a few 7204 NPE-225s serving as a PPPoA termination platforms
running 12.2(16)B2 (Enterprise Plus) to the very latest and greatest
12.4(3) or 12.4(2)T1 advanced enterprise image results in not all
connected customers being stable.
On one NPE-225 the amount of expected users is above 180 but "sh
caller sum" indicates that the amount of sessions connected is something
between 64 .. 71 .. 75 .. 84 .. 71 .. 64 .. 61 .. 58 .. 64". Meanwhile the
CPU is maxed out with 60% of time going to "VTEMPLATE Backgr", 5% to
"PPP Events" and 1.5% to "PPPoA Manager".
On another the number jumps between 140 and 155 but is still about 25
users too little.
All user configuration (full acl, acl name, framed routes, addressing, etc.)
is obtained from RADIUS.
Configuration is as follows:
virtual-profile if-needed
virtual-profile virtual-template 2
virtual-template 2 pre-clone 200
vc-class atm dsl-in
ubr 8192
inarp 1
encapsulation aal5mux ppp Virtual-Template2
transmit-priority 4
create on-demand
interface ATM2/0
mtu 1500
no ip address
no ip redirects
no ip proxy-arp
load-interval 30
class-int dsl-in
atm sonet stm-1
atm pvc 5 0 5 qsaal
atm pvc 16 0 16 ilmi
no atm auto-configuration
no atm ilmi-keepalive
no atm address-registration
no atm ilmi-enable
interface Virtual-Template2
description PPPoA/E customer
ip unnumbered Loopback0
no ip redirects
no ip proxy-arp
ip pim sparse-mode
no logging event link-status
no snmp trap link-status
ppp authentication chap pap callin AAA-DSL
ppp authorization AAA-DSL
I enabled "debug pppatm error" and "debug pppatm event" and stumbled on
messages like the following:
Aug 30 20:04:25.333: ATM1/0: PPPoA data path is not set for VC 5/180
Aug 30 20:04:25.693: ATM1/0: PPPoA data path is not set for VC 5/224
Aug 30 20:04:25.881: ATM1/0: PPPoA data path is not set for VC 5/179
Aug 30 20:04:26.253: ATM1/0: PPPoA data path is not set for VC 5/94
Aug 30 20:04:26.661: ATM1/0: PPPoA data path is not set for VC 5/149
Aug 30 20:04:26.857: ATM1/0: PPPoA data path is not set for VC 4/207
Aug 30 20:04:27.609: ATM1/0: PPPoA data path is not set for VC 6/65
Aug 30 20:04:42.965: ATM1/0: PPPoA data path is not set for VC 4/252
Aug 30 20:06:14.157: ATM1/0: PPPoA data path is not set for VC 5/162
Aug 30 20:08:01.021: ATM1/0: PPPoA data path is not set for VC 5/185
Aug 30 20:10:07.329: ATM1/0: PPPoA data path is not set for VC 5/185
Using Bug Tool I found CSCeh18155 which is supposed to be fixed in
12.4(1.8)T which I would assume means it is fixed in 12.4(2)T1..
Other interesting "sort of similar" bugs include CSCeh62257 as all the
interfaces created are of type (full) interface and not sub-interface
("sh vt") although this one is supposed to be fixed in 12.4(1.8)T too.
As majority of users (99%) use CHAP for authentication, so it wouldn't
be CSCin95447 either.
Has anyone else stumbled on something similar while upgrading old
routers?
// kaj
|
