LH Storage Node Failure Rate

[Steve McGee]

I've got a 2 node 10.8 virtualization SAN cluster and I'm curious if there are any known failure rates for a complete node failure, or anyone with experience of such?

My concern arises from the desired use of a non-network raid volume and the possibility of loss of data due to a node failure.

Anyone with any experience with this?
Thanks,
Steve

[Patrick Terlisten]

Hello Steve,

currently I can't report any complete node failures. Sure, some disks or PSUs are died, but no node was brought down by a FRU.

Regards,
Patrick

[Bryan McMullan]

I've run about 65 total LHN/HP nodes over the past 4 years and I've had two systems fail completely in that time. One was a raid controller failure on an old NSM160, and the second was a DL320s that had a bad mobo. In both cases, there was no outage and the issue was resolved without a production impact.

With the likely hood of drive failure, which has happened many times over the years... I would caution heavily against not using Network RAID. You'll have issues performing patches and upgrades as well as any reboot or storage server service recycle will cause your volumes to go offline.