Tom,

The disks could be kept busy.   Perhaps some checking would help to figure this 
out.

o find out the (dev_info_t *) dip for the two sata disks, and dump out their 
devi_pm_components
   from mdb, look for pmc_numlevels, it should be >1.  - most likely this is 
not an issue, but just to   
   be sure.

o run dtrace on both
        pm_busy_component(dev_info_t *dip, int cmpt)
and  pm_idle_component(dev_info_t *dip, int cmpt)
calls for the "dip" that matches the two sata disks' dip.

If the sata disks were idle'ed for too short a period (i.e. <60s), they won't
be powered down.  Then you might want to dtrace its strategy(9e)
routine to figure out what thread was keeping them busy.  There might be
a kernel thread constantly polling for the SMART data...

-jane
--
This message posted from opensolaris.org

Reply via email to