...just wanted to do some routine maintenance on my colo server, starting off by a md RAID scrub and checking whether SMART values were still fine and all. +Maik Zumstrull
may know where this is heading - he filed https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=700975
quite a while ago.TL;DR: 4 out of 6 drives in an active server lost.
Ok, easy to force reassembly once you know what you're looking for and in case the event counters are not too far off (...some disks being attached to a different controller), but still a hassle and a forced reboot that I'd rather have skipped.
Apparently there's a firmware "fix" out that removes some functionality of the Marvell 9230 and also may or may not limit PCIe speed to 2.5GBit x2 (instead of 5.0GBit x2), but I'm hesitant to deploy this remotely.
Thanks +Marvell Semiconductor
for reminding me why there's more expensive vendors out there...