Solaris 9 on a starcat class machine.
During SAN events the machine would panic due to a dangling timeout from the
ssd target driver thus..
unix: mutex_enter (0x18)
ssd: ssd_restore_throttle+8 (0x3003ea99a60)
genunix: callout_execute+0x90 (0x3000134e000)
genunix: taskq_thread+0x188 (0x300014238c8)
unix: thread_start+4 ()
The sd_lun structure was full of zeros and had been removed from the ssd_state
structure. This was assumed to be bugid 4758190 until it was realised that they
had the latest sd/ssd patches and it should have been fixed.
some diagnosis lead to this bug showing that the same thing can happen in ssd_unit_attach() when that fails.
tim
xxxxx@xxxxx.com 2005-05-17 14:14:53 GMT