cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
Announcements
Cisco Secure Email Support Community

Product Support Talos Support Cisco Support Reference + Current Release
Gateway Reputation Lookup Open a support case Secure Email Guided Setup
Gateway: 14.0.2-020
Cloud Gateway Email Status Portal Support & Downloads docs.ces.cisco.com
Email and Web Manager: 14.1.0-239
Email and Web Manager Web & Email Reputation Worldwide Contacts Product Naming Quick Reference
Reporting Plug-in: 1.1.0.136
Encryption Bug Search
Encryption Plug-in: 1.2.1.167
Cloud Mailbox Notification Service
Outlook Add-in(s): More info

1351
Views
0
Helpful
1
Replies
KonradW
Beginner

Cisco Ironport ESA cluster with different type of machines - how to load traffic ?

Hi,

 

I have very strange question but I hope that some one will be able to help me here.

On the beginning my Cluster Environment composed from 2x Ironport ESA C170. After two years I have decided to extend Cluster with the next two Systems, but CISCO has offered me much stronger new Systems 2x Ironport ESA C190.

 

At the Moment my cluster composed with:

2x C170

2x C190

These 4 machines are running in the configuration Cluster.

 

When I have ordered machines, CISCO told me that both Systems are compatible - in did, they are compatible till Moment when System starting to be overloaded. As I see, C170 is much slower than C190 but using the same configuration as C190.

When we are sending a lot of e-mails 2XC170 are running in conservation mode and they are not available, and 2xC190 are also running in conservation mode but I`m able to Login with ssh.

 

Now I would like to solve Performance Problem, to do it I have decided to build additional two VMs (8cores + 32GB RAM each one) and join them to the current cluster. What will happend if I will join that two VMs to the current "mixi dixi, super duper" Cisco Cluster ?

 

Will such a 6 nodes Cluster work ?

 

2x SLOW C170     (Hardware appliance)

2x MEDIUM C190    (Hardware appliance)

2x FAST VMs      (VMware appliance)

what will happend if I will send traffic with the same MX prio to all machines at the same time.

 

I`m really frustrated with CISCO Support, because each engineer has different opinion. Maybe some here has some better experience with Cluster sizing...

 

 

Thanks in advance for any hints.

 

 

Cheers

Konrad

 

1 REPLY 1
dmccabej
Cisco Employee

Hello,

 

There's not really any way to comment on what will happen if you add two new devices into the cluster without knowing exactly what is causing the current performance degradation. However, as far as an acceptable configuration goes, you can cluster up to 20 devices. So, you should have no issues running a 6 node cluster. 

 

For the MX question, the answer to that will depend on how the sending server decides to choose the next-hop. At that point, you're using DNS round-robin and hoping that it will equally balance out your email traffic, when in reality that's not normally the case. The ideal recommendation would be to add a load balancer into your network for true traffic dispersion across the devices. If you're on a budget though, DNS round-robin will still help ...Just not as much.

 

Thanks!

-Dennis M.

Create
Recognize Your Peers
Content for Community-Ad