My main production environment is in a Dell M1000e blade chassis with M6220 switches in Fabrics B and C. Fabric B is dedicated to vMotion traffic, and Fabric C is dedicated to iSCSi traffic and connected directly to our SAN. None of the switches in either Fabric are connected to the rest of the network, so I am unable to monitor them, and I can only manage the configs through the Dell Chassis Management Controller.

When our consultants set this up two years ago, they configured stacking as follows: C1=C2 B1=B2

I'd like to change the stacking configuration so it is looped in the following way: B1=C1=C2=B2

If I make this change and merge the two stacks, will it change the configs on the switches?


This can be done, see the page 15 of the Dell white paper on the subject here: http://www.dell.com/downloads/global/products/pwcnt/en/pwcnt_stacking_switches.pdf

Specifically it says:

Merging Two Operational Stacks

The recommended procedure for merging two operational stacks is as follows:

  1. Always power off all units in one stack before connecting to another stack.
  2. Add the units as a group by unplugging one stacking cable in the operational stack and physically connecting all unpowered units.
  3. Completely cable the stacking connections, making sure the redundant link is also in place.

Two operational stacks can also be merged by reconnecting stack cables without powering down units in either stack. Connecting a powered up standalone unit to an existing stack leads to same behavior as when merging two operational stacks. In such cases, Manager reelection is done and the Manager with the higher MAC address wins the election. The losing stack manager resets itself and all its member units. After the reset, all the losing stack members join the winning stack to form a single stack. The winning stack remains functional through the merge process. If the stack merge is performed in this way, then it is strongly recommended that the user set the admin management preference of the desired winner stack manager to a higher value than the stack manager that should lose the election.

Also, in your above proposed cabling you are not completing a loop with your stacking cables. You should always aim to do this, even though it is not a requirement. See page 7 of the Dell m6220 stacking use case: http://i.dell.com/sites/content/business/solutions/whitepapers/zh/Documents/powerconnect-m6220-switch_cn.pdf


M6220 switches should be stacked together by connecting a stacking cable from stack port xg2 of one switch to stack port xg1 of the next sw itch. The stacking cable is included with the stacking module. Stack port xg2 of the fi nal switch in the stack should be connected to stack port xg1 of the first switch, completing the stack loop. A loop topology is not strictly required for a stack but it is recommended, as it is the most robust stacking topology and has seen the most time in production use.

| improve this answer | |

The main part about this is the switch IDs for the devices. If you have two switch stacks which BOTH have units 1 and 2 your interfaces are named 1/g1, 1/g2, 2/g1 etc. When they are combined into one stack the configuration of at least two will change.


  1. Do backups of the configuration
  2. Get a service window :-)
  3. Shutdown two units, and join them to the stack
  4. Paste the configuration for the "new units" 3/g1-24, and 4/g1-24 etc.

TL;DR interface numbers follow switch id and switches cannot automatically merge two stack configs

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.