Skip to content

Comment on HPE Drive fail at 32,768 hours without firmware updateparent

Comments

We have a cluster of four nodes that were all setup and brought online within hours of each other. The entire cluster could blow up within a couple hours if not patched.

I guess cluster nodes should be scheduled to be taken down for random amounts of time so that they fail in sequence more gracefully.

Have an “off on weekends” node. And 24/7 nodes.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.