Firmware management?

Y0s

New Member
Feb 25, 2021
9
2
3
Hi,

How is everyone managing firmware for (heterogeneous) fleets of servers? I've done individual servers via IDRAC or ILO GUIs but there must be a better way? Is it possible to curate a local firmware repo then command all machines auto-update?

Thanks!
 

Y0s

New Member
Feb 25, 2021
9
2
3
Thanks for the overview. Is there a vendor independent solution? Has anyone used Redfish (or a tool built using the API)?
 

tssrshot

Member
Mar 18, 2015
54
7
8
Omaha, NE
This Is something my team is actually going to do a "platform sprint" on, this coming week.

Initial ideas are to try using Ansible/AWX to manage them via Playbooks, etc. AND Redfish. We're in between Dell, HPE, and Supermicro at the Datacenter AND the Edge. dell/redfish-ansible-module

We do some various form of the same thing with OneView, OpenManage, and individually on Supermicro...so this would be a much better solution for us, as well. Open to those ideas, folks.
 
  • Like
Reactions: Y0s

i386

Well-Known Member
Mar 18, 2016
2,453
675
113
31
Germany
One scare was enough for me, I wouldn't update multiple system simultaneously :D
(Bios update went wrong and the system didn't boot and could not be recovered with the super rom method. At the end supermicro helped with an undocumented command in sum that allows to force flash the bios remotely)
 
  • Like
Reactions: Y0s

Y0s

New Member
Feb 25, 2021
9
2
3
Good point, but I was assuming updates would follow the usual test/slow roll out practice.

Looks like some businesses are using fwupd to manage firmware?

One scare was enough for me, I wouldn't update multiple system simultaneously :D
(Bios update went wrong and the system didn't boot and could not be recovered with the super rom method. At the end supermicro helped with an undocumented command in sum that allows to force flash the bios remotely)
 

tssrshot

Member
Mar 18, 2015
54
7
8
Omaha, NE
Agreed. I think the initial concept, the Team here, is working from...is like how OpenShift or Ansible rolls-out...not all at once, but one at a time, and stop on failure.