UCS Server C220 (HDD Amber light Blinking)

[kick534]

Hello,

Am able ping the CIMC IP Address

But am unable to get the Web Gui or not even able to SSH the CIMC IP Address (Please advise ).

physically i can see that one HDD  amber light blinking in UCS server SUB.

1) I try to ready some already solved discussion that restarting the server will solve the issue for limited time.

2) But in my case one HDD is faulty so how should i proceed.???

3) Suppose if i change the hard disk with new one it will impact to production network or not ??? (What will happen with the data already stored in faulty HDD) it sync from cluster 

4) we have PUB in one chassis and SUB in one chassis ( HDD problem in SUB)

5) Or first i have to physically power down the UCS server and once i get the CIMC i can login the server after that i can change the HDD.

[Kirk J]

Greetings.

Chances are you are hitting one of the older CIMC mem leak issues.

To get out of that state, you need to power the server down, and actually remove the power cords for 30 sec, which is also known as a hard power reset.  This will force a fresh boot of the CIMC.

If you do power down the server, make sure you gracefully shut down the guestVMs first (I think the console of the pub/subs have a cli login, and option to power/shutdown), and then power off the ESXi host from the ESXi vsphere client/webapp.

To avoid the CIMC mem leak long term, you will want to upgrade to a recent version of the HUU that your C220M3 supports.

If you do, please check the UCS HCL tool to make sure your ESXi drivers get updates to match the firmware.

The UCS HCL tool is at https://ucshcltool.cloudapps.cisco.com/public/#

If your server is actually a BE6k, or similar config, you typically had a raid1 mirror OS boot volume, and a raid 5 datastore of 3 or more drives.

If the other drives in the raid volume are healthy, then replacing a single drive in a raid 1,5,6,10 type raid will not have a data impact other than I/O will be slowed down during raid rebuild.

That's something you should really open a TAC case for if you have warranty/contract active on it still.

Kirk...