Solved

Enabling NIC Teaming while servers in production - OK?

Posted on 2013-10-30
13
889 Views
Last Modified: 2013-10-31
Hello,

I recently discovered one of my clients had their main ESXi host plugged into a 10/100 switch.  I moved it to a gig switch and things are happy happy now. (ESXi 5.1)

I have 3 additional NICS on the server I'd like to enable NIC teaming with.  I have cabled the other 3 nics and plan on using the default NIC Teaming settings:

Load balancing: checked - route based on originating virtual port ID
Network fail over detection: not checked
Notify switches: checked - Yes
Fallback: checked - Yes
Override vswitch failover order: checked.

Move the 3 unused adapters into the active adapters group and apply.

My question is this:  Can this be done w/o having to shutdown guests?  It is VERY difficult to schedule downtime and if I can do this during business hours, that would be ideal.

I've read that I have to reboot the individual VM's to get them to 'load balance' themselves. It will be much easier to schedule individual reboots rather than their entire network.

Only 1 vSwitch, flat network (no vlans).

Thanks!
0
Comment
Question by:GDavis193
  • 5
  • 4
  • 4
13 Comments
 
LVL 117

Expert Comment

by:Andrew Hancock (VMware vExpert / EE MVE)
Comment Utility
I would not recommend to complete any network operations during production.

You will find making network changes will disrupt networking for a few seconds, and a few pings will be dropped.

and if you get it wrong, networking will be completely disrupted.

So best to schedule for Out of Core Hours.
0
 
LVL 9

Expert Comment

by:dipopo
Comment Utility
Just a quick input around the VMware statement.

"Uplinks added to an existing NIC Team will be active but un-usable until added to the LACP group on the physical switch"

I think removing an uplink causes disruptions but not the other way. Having said that "Please perform this out of hours"

I will monitor this closely, interesting to find out if any disruptions occur from adding.
0
 
LVL 9

Expert Comment

by:dipopo
Comment Utility
Also will need to change the load balancing from the current one to IP Hash if using LACP, this will guarantee the uplinks are evenly utilized for load balancing.

Originating port ID and MAC hash just assigns vNIC's arbitrarily. Else use physical NIC load as this looks at the load per NIC. In fact I feel this new policy is much better as it does not stipulate a need for LACP.
0
 
LVL 117

Expert Comment

by:Andrew Hancock (VMware vExpert / EE MVE)
Comment Utility
If you practice ITIL or Change Management, Change Control in your organization, no changes to a server should be conducted in Core Hours!

That would be our recommendations.
0
 
LVL 9

Expert Comment

by:dipopo
Comment Utility
Of course possibly service impacting changes are best planned out-of-hours as I concurred in my response above.

But coming from a purely technical "What will happen" I think it is best to be able to ascertain how such a change is expected to behave, hence my input.
0
 

Author Comment

by:GDavis193
Comment Utility
I appreciate the response but my question was not asked to determine when the best time to conduct this change but whether or not anyone has experience with this and can let me know if it can/has been done during production hours?  

Like increasing the size of a disk volume during production hours... it's best to schedule this after hours but there are times when the needs to be done asap.  It's a safe enough operation to be done anytime.

Again, not looking at best time to do this (everything should be done after hours, we all know this) but whether or not anyone has performed this on a live system.  

Thank you.
0
Control application downtime with dependency maps

Visualize the interdependencies between application components better with Applications Manager's automated application discovery and dependency mapping feature. Resolve performance issues faster by quickly isolating problematic components.

 
LVL 117

Expert Comment

by:Andrew Hancock (VMware vExpert / EE MVE)
Comment Utility
As I posted.....in my first post...when we have made these changes.... and it also depends how active and saturated your network links are

"You will find making network changes will disrupt networking for a few seconds, and a few pings will be dropped." Whilst routing tables converge.

If you get the configuration completely wrong, you will experience a total outage..

If you have Outlook Clients, they will get disconnected.
0
 

Author Comment

by:GDavis193
Comment Utility
You did not make it clear you've personally made those changes - you made what appeared (to me) to be a generic statement of:

"You will find making network changes will disrupt networking for a few seconds..."

"If you get the configuration completely wrong, you will experience a total outage.."

So to clarify,  are the basic settings outlined in my initial post are safe to do with knowledge of the brief network outage?  My main concern is that for some odd non-documented reason my management IP would be hosed or something...

This WILL be done at the end of the business day with only 3-4 users in the office.

Thanks in advance for the clarification.
0
 
LVL 9

Expert Comment

by:dipopo
Comment Utility
If referring to these settings:

Load balancing: checked - route based on originating virtual port ID [It will load balance but VM's will be assigned to a NIC when they start, it does not take into consideration physical NIC load, in my opinion route based on physical NIC load is best.]

Network fail over detection: not checked

Notify switches: checked - Yes [OK, unless when running MS NLB in Unicast]

Fallback: checked - Yes [OK, but a problem if flapping occurs]

Override vswitch failover order: checked.

In my opinion the setting is OK
0
 
LVL 117

Accepted Solution

by:
Andrew Hancock (VMware vExpert / EE MVE) earned 500 total points
Comment Utility
We've made these changes on behalf of clients, when we have advised them accordingly.

Expect a few seconds loss of pings, and Outlook Clients to disconnect, and Terminal Server connects or PCoIP clients to disconnect.

Depends on your physical network switches, and current activity on the switches.

Looks okay, as per disclaimer above.
0
 

Author Comment

by:GDavis193
Comment Utility
Outstanding!

Yes, I am ok with the brief network outage.

Thank you!
0
 

Author Closing Comment

by:GDavis193
Comment Utility
The guy is a little rough around the edges but I received the answer i was looking for.

A+ for technical assistance, C (wont be grading this) for the perceived forum 'attitude' :)
0
 
LVL 117

Expert Comment

by:Andrew Hancock (VMware vExpert / EE MVE)
Comment Utility
Apologies if I've offended, it was not intentional.

Good Luck with your network changes.
0

Featured Post

How to improve team productivity

Quip adds documents, spreadsheets, and tasklists to your Slack experience
- Elevate ideas to Quip docs
- Share Quip docs in Slack
- Get notified of changes to your docs
- Available on iOS/Android/Desktop/Web
- Online/Offline

Join & Write a Comment

Suggested Solutions

Title # Comments Views Activity
Vmware Affinity Rules 10 51
OSPF Design NSSA 5 30
VMware ESXi 5.5 - Remote Latency via VPN 2 29
Vcenter 6.5 install Error 8 50
It Is not possible to enable LLDP in vSwitch(at least is not supported by VMware), so in this article we will enable this, and also go trough how to enabled CDP and how to get this information in vSwitches and also in vDS.
Last article we focus in how to VMware: How to create and use VMs TAGs ā€“ Part 1 so before follow this article and perform the next tasks, you should read the first article how to create the TAG before using them in Veeam Backup Jobs.
Teach the user how to configure vSphere clusters to support the VMware FT feature Open vSphere Web Client: Verify vSphere HA is enabled: Verify netowrking for vMotion and FT Logging is in place or create it: Turn On FT for a virtual machine: Verify ā€¦
This video gives you a great overview about bandwidth monitoring with SNMP and WMI with our network monitoring solution PRTG Network Monitor (https://www.paessler.com/prtg). If you're looking for how to monitor bandwidth using netflow or packet sā€¦

743 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question

Need Help in Real-Time?

Connect with top rated Experts

7 Experts available now in Live!

Get 1:1 Help Now