04-14-2005 06:38 PM
Hello,
I am a new comer, lack of cisco knowledge. =P
I have a server with 2 network path access, like this
server <==seg1=====> router <==seg3===> client
\=====seg2=========/
the router has been config to transmit packet in both path, however, in a case that seg1 becomes unstable (problems on cables, hubs, switch, etc), it does not forward the packets to seg2, what should I do?
Thanks.
Joshua.
04-17-2005 03:29 AM
If I understand your drawing, there are two segments connecting the server and two interfaces on the router. So there are two subnets connecting the server and the router. One question would be what address are end stations using to access the server. If end stations use the IP address of interface 1 on the server, then the router will use its interface on the subnet connected to interface 1. If there is a problem (cable, hub, switch etc) then there is not much that the router can do. The router will not know that forwarding to interface 2 will reach the server, and the router can only try forwarding on interface 1.
I would think that the optimum solution would be to use a logical interface on the server (loopback) and to run some routing protocol between the server and the router so that the router will know that it can reach the logical interface via either of the two interfaces. An advantage of this is that the routing protocol will detect if there is a problem and converge sending all of the traffic on the surviving interface. If you do not want (or can not run) a routing protocol between the router and the server, then you could configure static routes on the router. There might be some disadvantage to this because it would not detect a cable problem between the switch and the server, or a problem on the switch.
HTH
Rick
04-18-2005 12:54 AM
How about if I set sever like this:
server
|
[vIPa]<==seg1=====> router1 <==seg2===\
\=====seg1====> router2 <===seg2====>[vIPb] - client
vIPa is a virtual IP act as default gateway of seg1, vIPb is a virtual IP act as default gateway of seg2.
the binding of the 2 vIP done by neg. of router1 router2 (I am not sure the name of that function), however, seems router1 always hold the binding of vIP, which router2 do not take up the vIP in case that router1 becomes unstable....
can I make some setting on these routers to simulate failover feature?
04-18-2005 04:52 AM
I am confused by this question. The first version of the question specified that the server had two interfaces and asked about the router getting to the server. This version of the question does not seem to have a second interface on the server and seems to be more focused on how the server (and the client) would get to the router. There is a feature of Cisco IOS called HSRP (Hot Standby Router Protocol) which addresses this function. If there are two routers on a common LAN, they can be configured with HSRP and will share a virtual address. The routers negotiate which one will be the active router and claim the virtual address. This virtual address is the default gateway of the end stations on that segment (for the server or for the client). If there is a problem on the active router, then the standby router will claim the virtual address and the end stations still will have an active default gateway.
Look on the Cisco site and you will find many references to HSRP which may be helpful.
HTH
Rick
04-18-2005 06:11 PM
This is another saturation. I just want to introduce some FT solution to my network, sorry if confusing.
What I mean is, although HSRP is on, it seems not going to perform the hot standby ability when one of the router become unstable or even go death, the other one do not take control of the vIP.
Is there any setting I should make?
04-19-2005 06:58 AM
If the question is to describe a completely different situation then I believe that HSRP on the router interfaces on the LAN should give you the redundancy that you are looking for. I am not sure why you say that it will not work. I have used this on a number of routers and my experience has been that if it is properly configured, and the lead router does fail, then the backup router does take over and the end stations do continue to have connectivity to remote destinations.
To answer your question about are there any settings that you should make, you should configure HSRP on both of the router interfaces on the LAN. At a minimum that would be standby 0 ip
There are other commands which are optional and which are documented in the HSRP documentation.
HTH
Rick
04-19-2005 06:36 PM
Yup, that was what I thought but it just doesn't work. I am sure that the vIP is working but the healthy router just not to take over the vIP sharply as the problem router become unstable.
May be it's due to the problem router is not completely death (but it is observable unstable, some packet lost in ping), I dont know is there any setting I can make to improve this take over action, since I am not experience on this...
04-19-2005 07:00 PM
Especially for HSRP unstable is one thing and dead is another. When running HSRP the routers send data packets to each other to determine which should be the active router and claim the vIP. If the router is up enough to send and receive these packets then the active router will remain active and claim the vIP - even if there is some packet loss in pings. And HSRP is not intended to change which router is active based on some ping loss. But if the active router really does stop functioning (dead) then the standby router will become active and the end stations will still have a working default gateway.
I wonder in your situation if one of these routers is more stable than the other one. If so you should probably configure priority in HSRP so that the more stable router will become the active router by having a higher priority.
HTH
Rick
04-19-2005 07:43 PM
How to configure that? Sorry that I do not have the experience.
BTW, I read some doc, and what idea I get is the priority setting is 1 way only. Means if I set the priority on router2 and this router2 become unstable, router 1 still cannot take up the vIP, right?
If so, it is not something I desire or I make the wrong understanding?
04-20-2005 04:56 AM
I am not sure what you mean when you talk about priority as 1 way only. If you set priority on a router (the command is standby 0 priority
I am not sure that you have a correct understanding of HSRP. You talk about a router being unstable but HSRP is not intended or designed to compensate for unstable. HSRP is designed to compensate for failure and it does that quite well.
HTH
Rick
04-21-2005 12:32 AM
server
|
[vIPa]<==seg1=====> router1 <==seg2===\
\=====seg1====> router2 <===seg2====>[vIPb] - client
Assume the problem is router become unstable only, but not death... (I trust HSRP can handle the total death case well)
what I mean is, if I set priority on router2 is slightly higher then router1, then it is likely that router2 will take the control of vIP if router1 has problem... however, in reverse, router1 would not take control of vIP router2 has problem, am I interpret right?
If I set both router1 and router2 has same priority, (actually the current status, as I haven't set both?), do you means I need set a "preempt" option on both router1 and router2 to let them take over the vIP in case of the opposite router has problem?
04-21-2005 05:20 AM
You need the preempt command to allow a router to take over as the active router based on a change in priority. If the active router fails the standby router will take over whether preempt is configured or not.
HTH
Rick
04-26-2005 01:13 AM
Thank you.
I need to study a bit about preempt options.. may have more questions later.
=D
05-03-2005 12:31 AM
How about if our router has enable the preempt option but the standby router is still unable take over when the active router becomes unstable....
Is there anything I miss? Or any things I need to setup?
05-05-2005 06:38 PM
Anyone have idea for this?
😛
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