[Bug 77376] Hard drive spin-down causes hang

Mark 3eq9v5q02 at sneakemail.com
Thu Dec 28 23:58:52 UTC 2006


Public bug reported:

Standard PC, NOT A LAPTOP, NOT A BATTERY ISSUE
Dapper 6.06 LTS Intel x86 32-bit with all updates applied

Implementing IDE hard drive spin-down in hdparm.conf causes strange
hangs in Dapper.  These hangs require forced reboot, sometimes full
power cycle.  Example, screen saver lockup, lockup at boot during
"loading hardware," etc.  Ctrl-alt-delete not a way out.

Hangs relate to changing the hard drive spin-down since none happened
before changes made.  Perfectly fine before that, ran for days at a
time.  Could be upstream Debian issue.  Guessing:  Linux is not properly
re-activating the drives on demand.

Not a physical jumper setting issue or motherboard misconfiguration.  I
know what I'm doing in those regards.  All is correctly set.  The drives
generally work.  The hdparm.conf settings might be up for debate.

Two IDE chains:
IDE1 = hard drive master, hard drive slave, identical drives
IDE2 = CDRW master, nothing else on this chain

/etc/hdparm.conf is the following

# IBM DTLA 307030 in udma4 mode (5 is not optimal, errors happen)
# acoustic 128 is "quiet" apm 255 is "off" for adv. power mgmt.
# IDE chain 1
/dev/hda {
	write_cache = on
	mult_sect_io = 16
        io32_support = 3
        dma = on
	transfer_mode = 68
	interrupt_unmask = on
	keep_settings_over_reset = off
	acoustic_management = 128
	spindown_time = 244
	apm = 255
}

/dev/hdb {
	write_cache = on
	mult_sect_io = 16	
        io32_support = 3
        dma = on
	transfer_mode = 68
	interrupt_unmask = on
	keep_settings_over_reset = off
	acoustic_management = 128
	spindown_time = 244
	apm = 255
}

# IDE chain 2
# HP CD-Writer Plus 9700i
/dev/hdc {
    quiet
    io32_support = 3
    dma = on
    transfer_mode = 61
    interrupt_unmask = on
    keep_settings_over_reset = off
}

** Affects: linux-source-2.6.15 (Ubuntu)
     Importance: Medium
         Status: Unconfirmed

-- 
Hard drive spin-down causes hang
https://launchpad.net/bugs/77376




More information about the kernel-bugs mailing list