We have a big linux box running on top of scalemp. The machine consists
physically of four nodes each with two LSI 9260 controllers. ScaleMP
makes it look like one big box. I.e. there are 8 controllers in total,
and also 8 SSDs connected to the onboard SATA controllers.
It is running CentOS 5.6 and the kernel version is 188.8.131.52-6 with
Scalemp's patches on top.
When I run smartctl on the SSDs, connected the onboard SATA controllers,
there are no problems.
When I run smartctl to check one of the disks behind the LSI raid
controllers the machine hangs cold. I.e :
% /opt/smartmontools/5.43/sbin/smartctl -H --device=megaraid,0 /dev/sdb
smartctl 5.43 2012-06-30 r3573 [x86_64-linux-184.108.40.206-6.vSMP] (local build)
Copyright (C) 2002-12 by Bruce Allen, http://smartmontools.sourceforge.net
and after that we have to powercycle it. Tried intially with version
5.41. Same result. Can I somehow enable debug to get more info ? Have
anybody run in to the same trouble ?