[c-nsp] Cisco 7500 High Availability

[Eric Kagan]

>    I've got a customer with some 75xx that have dual RSP4s in them. 
> We've got identical dram/flash/12.0.32S code on them and I've started 
> fooling with the high availability stuff. I've received 
> feedback from a 
> couple people on NANOG that this doesn't work very well - the four to 
> eight minute recovery RPR works, but the faster modes do not behave.
> 
>   Can anyone comment on this? I'd like to make it go for this 
> guy, but I 
> don't want to spend forty hours learning all of the failure 
> modes for a 
> software product that works better on other hardware platforms ...
> 
> 

With the IOS and config below we see < 1 min failover (some as low as 15-30
seconds).  We have had an RSP crash and the other kicks in.  By the time we
figure out what seems "weird", its already passing traffic.

One thing I will mention, we had no luck with this in the lab !  If we tried
to "pull" an RSP to test it, the router would crash and force us to power
cycle to get it back up.  We just kind of went with it and seems to be ok.

We usually schedule a maintenance window to reboot it after a crash, but at
least we are not down in the middle of the day.

Eric


c7507#show ver
Cisco Internetwork Operating System Software
IOS (tm) RSP Software (RSP-PV-M), Version 12.0(29)S, RELEASE SOFTWARE (fc1)
System image file is "slot0:rsp-pv-mz.120-29.S.bin"
cisco RSP4 (R5000) processor with 262144K/2072K bytes of memory.
Slave in slot 2 is running Cisco Internetwork Operating System Software
IOS (tm) RSP Software (RSP-PV-M), Version 12.0(29)S, RELEASE SOFTWARE (fc1)
Slave: Loaded from system
Slave: cisco RSP4 (R5000) processor with 262144K bytes of memory.


redundancy
 no keepalive-enable
 mode rpr-plus
hw-module slot 2 image slot0:rsp-pv-mz.120-29.S.bin
hw-module slot 3 image slot0:rsp-pv-mz.120-29.S.bin