Re: With SAN should you still use internal disk for vios?

SAN Casters up? Yep. Know of two of them.

Both Lost everything, causing a total system rebuild. We consultants call that: "Profitable."

Both were customer's fault really. One had server and SAN plugged into a UPS that after about 5 days had a fight with the server and SAN. Server and SAN would see voltage go down, so they raise current. (Ohm's law, not Ohm's suggestion!) UPS would see current go up and push a little harder, raising voltage. Server and SAN react accordingly. Eventually this went from 'up a little' 'down a little' to swinging more and more and eventually the UPS simply shut down hard. Customer then chose to leave the entire setup powered off for a week. This let the cache batteries in the SAN go flat and therefore all data on all volumes was lost.

UPS was huge and easily able to handle the load, even with the server and SAN it was under 20% load. Turns out the IBM approved solution was to plug in two 1,500 watt heaters to the UPS and turn them on max. Pure resistance loads bringing the load up to a point where the voltage didn't fluctuate.

Not particularly efficient of course but it worked. (and no, no longer in play!)

Second customer had misunderstood power cables from Blade Center and didn't have them split correctly between A and B power. When one side was taken down for maintenance so did entire blade center. Somehow that smoked the SAN as well. Complete rebuild again.

- DrF

On 8/22/2022 9:27 AM, Rob Berendt wrote:

Of course with two VIOS lpars (for redundancy) and each running mirrored drives you end up with 4 NVMe drives. Just to run VIOS.

How much time does it save you to be able to get VIOS running while you are waiting on SAN? At best it probably keeps you from breathing down the storage guy's neck while you are waiting.

Has anyone on this list actually been in this situation where the SAN went casters up?

I suppose having some device which could 'ping' FC might tell you that it's the switch vs the SAN.


Rob Berendt