Solved: Re: two active DLSw remote-peers

fmasson · ‎03-13-2003

Hi,

having two active dlsw remote-peers from a remote side to different CIP routers (both connected to the same mainframe), which wolud be the best solution for holding all dlsw circuits only on one particular dlsw peer? dlsw remote-peer cost doesn't work the way i'd like, cause both peers transport circuits at the same time.

dixho · ‎04-07-2003

Based on the information provided, I do not see a reason why the circuits load balance between two routers. I think that you need to open a TAC case for further troubleshooting.

View solution in original post

dixho · ‎03-13-2003

How about backup peer? You define the CIP routers which should not handle any traffic as a backup peer of the primary CIP router. The only drawback is that you lose connectivity if the CIP fails.

Going back to the DLSw cost. DLSw cost should be a solution to me. How do you define the DLSw cost? Do you define it dlsw local-peer statement? Or do you define in the dlsw remote-peer statement? Show dlsw cap on the remote router is useful.

dixho · ‎03-13-2003

I forgot to ask you if you configure dlsw timer explorer-wait-time on the remote router. You need to increase the time 100. Please go to the following URL:

http://www.cisco.com/warp/public/cc/pd/ibsw/ibdlsw/tech/dls3_rg.htm

Look at the section on "Controlling Peer Selection."

fmasson · ‎03-14-2003

i only can define it on the remote-peer, cause it should not affect all remote-peer connections! dlsw cap shows the same cost on each remote-peer, so i conldn't really understood what cost setting on remote-peer should do?

but what about peer weight?

dixho · ‎03-14-2003

DLSw cost should work on dlsw remote-peer statement. Please post the dlsw remote-peer statement and show dlsw cap.

Peer weight is used for load balancing. In other words, how does DLSw to load balance circuits on 2 or more remote peer with the same cost.

fmasson · ‎03-14-2003

dlsw local-peer peer-id 10.1.1.77 group 254 keepalive 10

dlsw remote-peer 0 tcp 10.2.0.225 cost 4 timeout 90

dlsw remote-peer 0 tcp 10.3.0.226 cost 2 timeout 90

DLSw: Capabilities for peer 10.2.0.225(2065)

vendor id (OUI) : '00C' (cisco)

version number : 2

release number : 0

init pacing window : 20

unsupported saps : none

num of tcp sessions : 1

loop prevent support : no

icanreach mac-exclusive : no

icanreach netbios-excl. : no

reachable mac addresses : none

reachable netbios names : none

V2 multicast capable : yes

DLSw multicast address : none

cisco version number : 1

peer group number : 99

peer cluster support : yes

border peer capable : no

peer cost : 3

biu-segment configured : no

UDP Unicast support : yes

Fast-switched HPR supp. : no

NetBIOS Namecache length : 15

local-ack configured : yes

priority configured : no

cisco RSVP support : no

configured ip address : 10.2.0.225

peer type : conf

version string :

DLSw: Capabilities for peer 10.3.0.226(2065)

vendor id (OUI) : '00C' (cisco)

version number : 2

release number : 0

init pacing window : 20

unsupported saps : none

num of tcp sessions : 1

loop prevent support : no

icanreach mac-exclusive : no

icanreach netbios-excl. : no

reachable mac addresses : none

reachable netbios names : none

V2 multicast capable : yes

DLSw multicast address : none

cisco version number : 1

peer group number : 254

peer cluster support : yes

border peer capable : no

peer cost : 3

biu-segment configured : no

UDP Unicast support : yes

Fast-switched HPR supp. : no

NetBIOS Namecache length : 15

local-ack configured : yes

priority configured : no

cisco RSVP support : no

configured ip address : 10.3.0.226

peer type : conf

version string :

dixho · ‎03-18-2003

it looks like a bug to me. Please provide the IOS version and platform. I wil recreate it in a lab.

fmasson · ‎03-31-2003

...sorry for that delay! The two local CPA Routers (7206VXR) are running 12.2(7b) Enterprise SNASW Feature Set and the two remote 2612 running IOS 12.1(12) IP Plus.

dixho · ‎04-01-2003

After further research, I find out that the cost on show dlsw cap only shows the one define in dlsw local-peer peer-id statement (i.e. the value receives in DLSw cap exchange). The value defined in dlsw remote-peer statement is not shown anywhere. However, Cisco routers has a preference on the value defined in dlsw remote-peer over the value receive in dlsw cap exchange. In other words, what you see in show dlsw cap is normal.

I use a 7204 with 12.2(7b) Enterprise SNASw feature set. I do not see the problem. I would like to make a comment on dlsw cost. Circuit load balance is determined on the router whose connected SNA device initiates the LLC2 connection. 99% of time, it means that the router on the PU2.0 controller side determines. From the description, I am not sure if you configure the DLSw cost on the 7200 or 3600.

Please send me the configuration of all router @ dixho@cisco.com

fmasson · ‎04-03-2003

Thanks for your further research, please find the dlsw parts enclosed:

Primary 7206 (CPA):

dlsw local-peer peer-id 10.3.0.226 promiscuous

dlsw remote-peer 0 tcp 10.1.1.77 cost 2

dlsw remote-peer 0 tcp 10.1.2.78 cost 2 backup-peer 10.1.1.77

Backup 7206 (CPA):

dlsw local-peer peer-id 10.2.0.225 promiscuous

dlsw remote-peer 0 tcp 10.1.1.77 cost 4

dlsw remote-peer 0 tcp 10.1.2.78 cost 4 backup-peer 10.1.1.77

Primary 2612 (Customer):

dlsw local-peer peer-id 10.1.1.77

dlsw remote-peer 0 tcp 10.2.0.225 cost 4

dlsw remote-peer 0 tcp 10.3.0.226 cost 2

Backup 2612 (Customer):

dlsw local-peer peer-id 10.1.2.78 promiscuous

dixho · ‎04-07-2003

Based on the information provided, I do not see a reason why the circuits load balance between two routers. I think that you need to open a TAC case for further troubleshooting.