disk spin down <summary>

[Peter S Galbraith]

[Bottom line: Can scsi-idle be used with swap?]

I wrote:

> I apologize if this isn't the proper forum.  I don't know if this
> is handled by the SCSI driver or not.  I use Linux pre-patch-2.0.31-3.
> 
> I have a spare disk I use exclusively for nightly backups, and 
> optionally seldom used swap.  I'd like to spin down the disk when not 
> in use (and unmounted) because of it's whinning sound.  

Pierre Turlais told me about scsi-idle:

Title:          scsi-idle - Turn off SCSI disks when idle (Take 2)
Description:    A small kernel patch which spins down SCSI-disks after a 
                period of inactivity. 
                scsi-idle can probably be considered to be BETA software now, 
                it works for me and half dozen other people, but YMMV.

I patched the Linux kernel and tried the program when no partitions were
mounted.  It worked but I got the following messages:

 kernel: SCSI host 0 channel 0 reset (pid 4623) timed out - trying harder
 kernel: SCSI bus is being reset for host 0 channel 0.
 kernel: (scsi0:-1:0) Reset device, active_scb 1
 kernel: scsi0: (targ -1/chan A) matching scb to (targ 1/chan A)
 last message repeated 2 times
 kernel: scsi0: Resetting current channel A
 kernel: scsi0: Channel reset, sequencer restarted
 kernel: (scsi0:1:0) Aborting scb 1

I assume all this is normal.

Then I enabled a swap partition on the spun-down SCSI disk and it spun-up
again to turn on swap.  I told scsi-idle to spin down the disk after 5
minutes of inactivity.  It spun down after 5 minute as I never swap much.

And I got these messages:

 kernel: scsi : aborting command due to timeout : pid 5539, scsi0, channel 0, id 1, lun 0 Start/Stop Unit 00 00 00 00 00 
 kernel: general protection: 0000
 kernel: CPU:    0
 kernel: EIP:    0010:[<00000004>]
 kernel: EFLAGS: 00010202
 kernel: eax: 00000000   ebx: 002fd018   ecx: 001d8090   edx: 03b40003
 kernel: esi: 00000297   edi: 00000000   ebp: 002f0018   esp: 001d5efc
 kernel: ds: 0018   es: 0018   fs: 002b   gs: 0000   ss: 0018
 kernel: Process swapper (pid: 0, process nr: 0, stackpage=001d42e4)
 kernel: Stack: 0019757f 002fd018 002fd018 002f0018 00000206 001d5f50 001960d0 002fd018 
 kernel:        00000003 002fd018 00197a36 002fd018 00040000 001fecc0 00000001 00112048 
 kernel:        00000001 ffffffff 00000001 00000001 001d5f6c 001feb30 00117907 001d5f6c 
 kernel: Call Trace: [scsi_abort+279/508] [scsi_times_out+72/304] [scsi_main_timeout+134/168] [timer_bh+248/820] [do_bottom_half+59/96] [handle_bottom_half+11/32] [schedule+436/648] 
 kernel:        [exit_notify+219/468] [do_exit+487/492] [die_if_kernel+695/704] 
 kernel: Code: d0 9b 00 f0 c3 e2 00 f0 d0 9b 00 f0 d0 9b 00 f0 54 ff 00 f0 
 kernel: Aiee, killing interrupt handler
 kernel: SCSI host 0 abort (pid 5539) timed out - resetting
 kernel: SCSI bus is being reset for host 0 channel 0.
 kernel: (scsi0:1:0) Abort_reset, scb flags 0x1, while idle, LASTPHASE = 0x1, SCSISIGI 0x0, SEQADDR 0x5, SSTAT0 0x5, SSTAT1 0xa
 kernel: (scsi0:1:0) Queueing an Abort SCB.
 kernel: (scsi0:1:0): Abort message sent.
 kernel: SCSI host 0 channel 0 reset (pid 5539) timed out - trying harder
 kernel: SCSI bus is being reset for host 0 channel 0.
 kernel: (scsi0:-1:0) Reset device, active_scb 0
 kernel: scsi0: (targ -1/chan A) matching scb to (targ 1/chan A)
 last message repeated 2 times
 kernel: scsi0: Resetting current channel A
 kernel: scsi0: Channel reset, sequencer restarted
 kernel: (scsi0:1:0) Aborting scb 0

Anything to worry about?
Or should I never enable a swap partition on a scsi-idle'd disk?

Thanks!
--
Peter Galbraith, research scientist             <galbraith at mixing.qc.dfo.ca>
Maurice-Lamontagne Institute, Department of Fisheries and Oceans Canada
P.O. Box 1000, Mont-Joli Qc, G5H 3Z4 Canada  418-775-0852 - FAX 418-775-0546