UCS system has lots of components in it, i.e Blade server, Rack server, Fabric Extender, Fabric Interconnect. All these are managed by the UCS manager. This document provides the steps to gracefully shutdown and power up the UCS system.
In an event when you need to shutdown the entire system for some activity follow the steps mentioned below to gracefully shut the UCS system. It is recommended to backup the UCS configuration before doing this activity as mentioned here. This document is applicable to all hardware and software version of UCS. Note that the information in this document was created from the devices in a specific lab environment. If your network is live, make sure that you understand the potential impact of any command.
Follow the steps to gracefully shutdown UCS system:
a) Graceful shutdown of servers: To shutdown the OS running in the server, UCSM sends ACPI signal to the server OS. The OS generally responds by gracefully shutting down; however this may not happen in certain cases. It is advisable to manually check if the OS has shutdown. Disassociating service profile from the servers will also shutdown the severs; however this is not advisable and is not graceful.
b) Disable management services: This step is optional and is intended only to lower the admin burden because of UCS system going down. Disable services like “Smart call Home” to stop UCS system from sending logs and information.
c) Decomission and Remove power to Chassis:Decomissoning is required only in case if you need to change the chassis. In other cases you can directly remove the power to chassis. This is done to stop the power to blade/rack server and Fabric Extender. There is nothing to be configured or saved from chassis end point.
d) Remove power to Secondary FI: This step is optional and is required only if you have two FI in failover. Removing power first from secondary FI prevents the system from doing failover, which is unnecessary here.
e) Remove power to Primary FI: As the last step remove the power to the primary FI. Note that there is no graceful shutdown method in case of FI, since it is not required.
The UCS system is now properly down and you can proceed with the maintenance activity. To bring the system back up follow the above steps in reverse order; i.e starting from the FI and going up to the servers.
The UCS system should reboot without any issue. However after rebooting if you find that the configuration is corrupt, follow the method to restore the configuration as described here. If after rebooting the UCS system gets stuck in boot prompt then follow the method to recover it as described here.
Hello,Please I need help with ASA L4-L7 service graph. I got a lot of faults and I can't troubleshoot that , I need your help please . I will try to put the screenshots here in order for you to figure out what's the issue . First of a...
Wondering if there is an escape sequence possible for switch credentials in discovery. All our switches use RADIUS and credentials to login are the form email@example.com unless the RADIUS server is unavailable and the switch has re...
I have 2 ACI fabrics. One in my production DC and the other in a DR facility. They are separate fabrics with separate controllers. I need to extend an epg EPG1/BD1 from production to DR. I will use OTV for LAN extension. Can’t use some of the ...
I need to be able to cancel the successor jobs within a stream after a job dependency has been aborted by a long running job event. I have a job event "Running Past Time Window-ABORTED" that will cancel a job if it is running past the time window. I do no...
Hello,I am registered to Managing Cisco Nexus 9000 Switches with Cisco DCNM. however I have difficulties doing the last lab:Provision VXLAN and Manage Switch with Cisco DCNMI am unable to do Step4 E, when I click save I got the error:The fabric DC wa...