procedure to turn off the complete ACI

[FedeSaraullo]

Hello, i need to know how shutdown the leaf, spine and apic in a  ACI structure. The procedure.

Thanks a lot.

[gmonroy]

Hello,

Could you be a little more specific in your intent for shutting off the entire ACI fabric? Is it for a move, where we may care about building the fabric back up with the databases intact? or perhaps you just want to power everything down for storage?

In either case, the following guides can be used an initial reference depending on your answer to the above:

To decommission switch nodes

To shut down an APIC

To Decommission/Replace an APIC

Cheers,

-Gabriel

[FedeSaraullo]

Hi Gabriel! Thanks for your help. We are testing ACI equipment and apic GUI. We need shut down sometimes

[FedeSaraullo]

Hi! So, do you have any idea of procedure to shutdown de aci? server, leaf and spine?

Thanks

[Cristian Munoz]

Hi Fede,

In general, you can gracefully shutdown APIC via CLI or GUI. As previously suggested (To shut down an APIC), click on System, In the submenu bar, click Controllers, right click on the APIC controller and the shutdown option is available.

Via SSH and issue the command> acidiag stop. These steps would be the easiest way. 

Hope this helps!

[wes_ward]

Yoyr reference about doing a shutdown on a APIC to move states "https://www.cisco.com/c/en/us/td/docs/switches/datacenter/aci/apic/sw/1-x/troubleshooting/b_APIC_Troubleshooting/b_APIC_Troubleshooting_appendix_010100.html
•Always decommission an APIC cluster controller before doing a power cycle, then add it back to the cluster. Failure to follow this guideline can corrupt the APIC cluster database shards residing on the cluster, which will require doing a wipe of the controller, then a cluster synchronization to restore a valid copy of the APIC cluster database from the other APIC controllers in the cluster."

Other docment states the following.

https://www.cisco.com/c/en/us/td/docs/switches/datacenter/aci/apic/sw/1-x/troubleshooting/b_APIC_Troubleshooting/b_APIC_Troubleshooting_chapter_0101.pdf

Shut down a single APIC controller as follows:

If possible, move APICs one at a time. As long as there are at least two APICs in the cluster online, there

is read/write access. If you need to relocate more than one APIC at a time, this results in one or no

remaining controllers online, and the fabric will go into a read-only mode when they are shutdown. During

this time there can be no policy changes including Endpoint moves (including Virtual Machine movement).

Once the APICs are shut down using the following procedure, relocate the controller, and power it back

up under the new rack. Then, confirm that the cluster health returns to fully fit status.

Note

1 In the menu bar, click System.

2 In the submenu bar, click Controllers.

3 Under Controllers, click the APIC node that you would like to reload, for example, apic1 (Node-1).

4 In the right window pane, at the top of the screen, click the General tab.

5 In the right window pane, at the top of the screen and under the tabs, click the ACTIONS pull-down menu.

6 Select Shutdown from the pull-down menu to immediately reload the APIC controller.

Which is correct?  I need to move my APICs from their current racks to new ones.  I would like to do so safely and successfully.

[wes_ward]

I should mention I am on version 2.1(2g).

[madnirafik]

Starting version 3.0

You can use GIR to switchover traffic gracefully:

https://www.cisco.com/c/en/us/td/docs/switches/datacenter/aci/apic/sw/1-x/aci-fundamentals/b_ACI-Fundamentals/b_ACI-Fundamentals_chapter_010011.html#id_54178