Design& Config help

[Shibu1978]

Dear All,

We need the following switching Network at our office.

1. Access-Layer switches are being connected to User Workstation.

2. Severs are being connected to Core-Switch + 2 more Switches ( Please see the Attached diagram)

Right now we plan to have redundancy in the network. we are going to have one more Core Switch in the network and all access-layer switches are planned to have fiber uplink to both switches.

Please clarrify me the following doubts.

1.Since we have many servers connected to current Core Switch ( cisco 4507) while planning redundancy (HSRP) how can we design the network? do we need to shift the servers to a different switch or we can still have these servers in the core switch and plan for redundancy?

2.Is there any better way we can achive the redundancy?

Thanks for all your help.

[ohassairi]

you can still have these servers in the core switch. just put them in one vlan and configure hsrp for the SVI of this vlan.

[Giuseppe Larosa]

Hello Shibu,

nice to see your project is going on.

1) the best solution is to have the servers on an access layer switch with two uplinks one to core1 and one to core2.

servers with only one NIC if connected only to core1 cannot take advantage of the presence of core2: if core1 fails they are just isolated.

However, some servers can be connected to a core switch.

Be aware that HSRP requires a dedicated IP address, the end result is that in each subnet three IP addresses are used by core1, core2, and HSRP VIP.

2)

for client vlans you could consider GLBP for better load balancing results (true only when multiple clients are present).

Consider GLBP only if you are able to connect each access layer switch with two uplinks to both cores as discussed in a previous post.

Hope to help

Giuseppe

[Shibu1978]

Dear Giuseppe,

As always thanks for your guidance.

Unfortunately we can not shift the severs from Core switch to Access layer Switches.I cannot even disturb the servers now at any cost. we are aslo going to have IP telephony in the network.Avaya is been selected as IP phones.

Hence here is the plan i propose to the management.

1.Buy one more Core Switch and connect fiber link to all access layer switches and L2 trunk between core switches.

2.Let the servers be in Core switch 1 and configure HSRP/VRRP between core switches.

3.Second Core Switch will have the same GIG port installed for server connection and configure accordingly and keep it ready. when Core-Sw1 fails we will unplug the cables from it and will connect to Core-Sw2.. does the solution work out? is that good solution?

Thanks for your reply.

Regards,

Shibu

[Giuseppe Larosa · Accepted solution]

Hello Shibu,

your action plan looks like reasonable.

final note:

investigate if servers can have a second NIC to be used for active/standby NIC teaming this would allow you to deploy cabling to standby NIC of each server.

be aware that different server vendors use different names and different teaming methods.

See this as a possible improvement.

Hope to help

Giuseppe

[Shibu1978]

Dear Giuseppe,

Thanks & sorry for the late written reply. Last two days i was not in the office.

Yes servers have been configured with Active/Active Teaming and it have been running for the last 5 years.

My proposal has been given to management but yet to get reply on that .

will bother u somedays later.. thanks for u r kind advice all the time.