US 12,353,550 B2
Snapshot-based boot volume for reliable server management and immutable reboot processes in storage processing units
David DeJong, Fremont, CA (US); Tobias Flitsch, San Jose, CA (US); and Siamak Nazari, Mountain View, CA (US)
Assigned to Nvidia Corporation, Santa Clara, CA (US)
Filed by NVIDIA CORPORATION, Santa Clara, CA (US)
Filed on Feb. 28, 2023, as Appl. No. 18/115,260.
Claims priority of provisional application 63/314,970, filed on Feb. 28, 2022.
Claims priority of provisional application 63/314,987, filed on Feb. 28, 2022.
Claims priority of provisional application 63/314,996, filed on Feb. 28, 2022.
Claims priority of provisional application 63/316,081, filed on Mar. 3, 2022.
Prior Publication US 2023/0273799 A1, Aug. 31, 2023
Int. Cl. G06F 9/4401 (2018.01); G06F 11/14 (2006.01); G06F 21/56 (2013.01)
CPC G06F 21/566 (2013.01) [G06F 9/4403 (2013.01); G06F 9/4416 (2013.01); G06F 11/1446 (2013.01); G06F 2221/033 (2013.01)] 13 Claims
OG exemplary drawing
 
1. A method for operating a storage system that includes a server, the method comprising:
storing an operating system image from a trusted source into a boot volume for the server, wherein the server contains a storage processing unit connected to control backend storage and provides the boot volume to the server and wherein storing the operating system image comprises the storage processing unit downloading the operating system image from a cloud-based management service for the storage system;
storing one or more component images from the trusted source in the boot volume, the components being selected according to a desired configuration of the storage system, wherein storing the component images comprises the storage processing unit downloading the component images from a cloud-based management service for the storage system;
after storing the component images, booting the server from the boot volume, the server installing components from the component images while booting;
after booting the server, taking a snapshot of the boot volume; and
configuring the storage system to promote the snapshot to be the boot volume before every reboot of the server, whereby the server always reboots to a known operable state.