11-15-2004 11:41 AM
ICMP and UDP still work outbound from the real servers, but TCP doesn't. Did something change? Should I go back to 4.1.1?
TIA
Simon
11-15-2004 11:59 AM
It was an asymetric routing issue, don't know why it hadn't happened under 4.1.1
Adding a load of specific routes on the right VLAN fixed it.
11-29-2004 02:26 AM
I have got the same problem.
So, I have found that solution.And,I see your comment in this forum.
But,I don't fix it.
Let me know the how and why of it.
Thank you.
11-29-2004 07:24 AM
Do you have a 'predictor forward' serverfarm / vserver?
Is the real server included in a server farm already?
Can you post you config?
11-29-2004 04:48 PM
Hi,
At First,Thank for your reply.
And, Your questions is YES.
I think that Our Config is OK.
The following is config.
module ContentSwitchingModule 6
vlan 10 server
ip address 192.168.16.223 255.255.255.128
alias 192.168.16.224 255.255.255.128
!
vlan 482 client
ip address 192.168.16.223 255.255.255.128
gateway 192.168.16.129
!
probe TEST-SERVICE http
request method head url /index.html
expect status 200
interval 20
retries 2
failed 30
!
serverfarm TEST-220
nat server
no nat client
failaction reassign
real 192.168.16.195
inservice
real 192.168.16.196
inservice
health retries 20 failed 120
!
serverfarm TEST-221
nat server
no nat client
failaction reassign
real 192.168.16.177
inservice
real 192.168.16.181
inservice
health retries 20 failed 120
!
serverfarm SERVER_DIRECT
nat server
no nat client
predictor forward
!
sticky 220 netmask 255.255.255.255 timeout 60
!
sticky 221 netmask 255.255.255.255 timeout 60
!
vserver DIRECT_ACCESS
virtual 192.168.16.128 255.255.255.128 any
serverfarm SERVER_DIRECT
persistent rebalance
inservice
!
vserver TEST-220-11000
virtual 192.168.16.220 tcp 11000
serverfarm TEST-220
sticky 60 group 220
persistent rebalance
inservice
!
vserver TEST-220-12001
virtual 192.168.16.220 udp 12001
serverfarm TEST-220
sticky 60 group 220
persistent rebalance
inservice
!
vserver TEST-221-11000
virtual 192.168.16.221 tcp 11000
serverfarm TEST-221
sticky 60 group 221
persistent rebalance
inservice
!
vserver TEST-221-12001
virtual 192.168.16.221 udp 12001
serverfarm TEST-221
sticky 60 group 221
persistent rebalance
inservice
!
diagnostic cns publish cisco.cns.device.diag_results
diagnostic cns subscribe cisco.cns.device.diag_commands
!
redundancy
mode rpr-plus
main-cpu
auto-sync running-config
auto-sync standard
!
vlan internal allocation policy ascending
!
vlan 10,20
!
vlan 50
name client
!
vlan 462,482
!
!
interface GigabitEthernet1/1
no ip address
speed 100
duplex full
switchport
switchport access vlan 10
switchport mode access
spanning-tree portfast
!
interface GigabitEthernet1/2
no ip address
speed 100
duplex full
switchport
switchport access vlan 10
switchport mode access
spanning-tree portfast
!
interface GigabitEthernet1/3
no ip address
speed 100
duplex full
switchport
switchport access vlan 10
switchport mode access
spanning-tree portfast
!
interface GigabitEthernet1/4
no ip address
speed 100
duplex full
switchport
switchport access vlan 10
switchport mode access
spanning-tree portfast
!
interface GigabitEthernet1/5
no ip address
speed 100
duplex full
switchport
switchport access vlan 482
switchport mode access
spanning-tree portfast
!
interface GigabitEthernet1/6
no ip address
speed 100
duplex full
switchport
switchport access vlan 482
switchport mode access
spanning-tree portfast
!
interface GigabitEthernet5/1
no ip address
switchport
switchport access vlan 462
switchport mode access
!
interface GigabitEthernet5/2
no ip address
switchport
switchport access vlan 482
switchport mode access
!
interface Vlan1
no ip address
!
interface Vlan462
ip address 172.6.46.254 255.255.255.252
!
interface Vlan482
ip address 192.168.16.129 255.255.255.128
ip route-cache flow
!
ip classless
ip route 0.0.0.0 0.0.0.0 172.6.46.253
no ip http server
!
11-29-2004 09:39 PM
HI,
the config looks more or less fine. please remove the NAT Server statement in the serverfarm SERVER_DIRECT. Normaly it shouldn't be a problem as there is no real mentioned where the CSM has to do the NATing for but I experienced strange things in the lab havig configured the server nat. They went away after I removed that command.
Could you do me the favour and take a sniffer trace in VLAN 482 and VLAN 10 when trying to connect to the servers directly and post them here?
Kind Regards,
Joerg
11-29-2004 11:28 PM
I had already removed the nat server in serverfarm SERVER_DIRECT.(no nat server)
But,couldn't solve it.
I had monitored my problem through EtherPeekNX.
So, I have captured file.
I'll attacted it.
This file was captured on port-channel262(the mux interface of CSM).
My problem is the following
Form client to virtual and real server.
I get OK on service including ICMP,UDP & TCP.
From real server to client.
OK on service including ICMP,UDP.
But,I cann't establish TCP connection because of TCP repeated Connect attempt.
02-28-2005 11:49 AM
Is there a way to control which client VLAN traffic sourced from a real will exit the CSM through? Other than static routed that is?
03-07-2005 08:10 PM
I don't have any config relating to question.
But,I had solved my problem.
After delete the following Config on CSM.
serverfarm SERVER_DIRECT
nat server
no nat client
predictor forward
!
vserver DIRECT_ACCESS
virtual 192.168.16.128 255.255.255.128 any
serverfarm SERVER_DIRECT
persistent rebalance
inservice
I think that don't need 'predictor foward' in bridged-mode.
03-08-2005 12:55 AM
You are right, in bridged mode no predictor forward is needed.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide