An old client's one-line note made me rethink all my test setups

Last month I had a client in Tacoma who kept losing data on a RAID 5 array. I'd swapped drives, rebuilt it twice, checked every log. After the third visit they left a sticky note on the server that just said "have you tried the other power supply?" I always assumed the redundant PSU was fine since the unit showed green. Turns out one rail was sagging under load, causing intermittent resets that corrupted the array. Took me 20 minutes to verify with a multimeter. Now I check both PSUs on any server job before touching the drives. Anyone else ever get a tip from a customer that made you feel like you'd been doing it wrong for years?

1 comments

1 Comment

barbaraw4723d ago

Learned that one the hard way too, a flaky PSU made a fresh RAID rebuild vanish overnight.