Automatic means for spinning down disks available?
Garrett Cooper
youshi10 at u.washington.edu
Mon Apr 9 00:33:18 UTC 2007
Jonathan Horne wrote:
> On Sunday 08 April 2007 18:11:51 Garrett Cooper wrote:
>
>> Gary Kline wrote:
>>
>>> On Sun, Apr 08, 2007 at 10:10:17PM +0400, Yuri Grebenkin wrote:
>>>
>>>> Just wonder if it's better for an HDD not to spindown at all.
>>>> Maybe it's safer to spin in peace than to park/launch?
>>>> What do you think?
>>>>
>>> My guess (really a SWAG) is that it's bettter to leave things
>>> just happily spinning, 24*7. In Nov, '99 a power off//on
>>> destryed my new (105-day-old) 9G SCSI drive. Off ffor fewer
>>> than five seconds, then a spike or two, and the drive went
>>> deadder than a decade-old corpse. Lost 10 months of files.
>>> ((Well, my tape backup had flubbed up.))
>>>
>>> Who would know??? I've heard both sides, and so far, just
>>> leaving drive spin seems slightly better.
>>>
>>> {Futureistic[?] idea: maybe a new drive can have a mode of
>>> Full-Operation and (slower) Spin. It wouldn't take more than
>>> a second to transition from the slow-spin to full-op mode.
>>> Open files, OS states, and whatever could be stored to RAM... .
>>>
>>> Any little old winemakers, er, diskmakers out there?
>>> }
>>>
>> Good point. The worst stress points during a disks life are at spin-up
>> from what I've read.
>>
>> Also, about the disk spinning at different speeds: many contemporary
>> disks have "acoustics" levels where you can adjust the speed on demand
>> (assuming you knew the hardware level instructions to send to the
>> controllers). Unfortunately I don't know those settings, so I can't say
>> what is and isn't possible.
>>
>> The only upside is at least all disk makers seem to be amalgamating into
>> either: Fujitsu, Hitachi, Quantum, Seagate, and WD, so figuring out the
>> standards shouldn't be *too* hard =).
>>
>> -Garrett
>>
>>
>>> gary-the-thrifty
>>>
>>>
>>>>> Hello again all,
>>>>> I was wondering if there was an automatic, and possibly timed means to
>>>>> spin down disks available in either ports or the base system, by
>>>>> chance. Just trying to cut down on energy use, and increase my disks'
>>>>> lives :). TIA,
>>>>> -Garrett
> personally, my solution for solving the "lower power consumption but still
> remotely available" issue, by configuring Wake On Lan. my web server is
> always on, so i just installed net/wakeonlan there. simple lines in crontab
> wake all the rest of my hosts each morning (after im gone to the office of
> course) for backups, and then they all power themselves back down about 2
> hours later. during the day, if i need to get to a system while im still
> remote, i just log into the webserver and wake it backup again.
>
> i would agree that the greatest stress on a disk might just be while its
> turning on from cold... but with the warranties that seagate is offering
> these days, i feel bold enough to power them off/on at least once a day.
Well, I feel the same but only about WD's drives. Seagate's newer drives
seem to die a lot more frequently than they used to (I've had 4 / 7
Seagate drives die on me in the past few months and 1/6 WD drives die on
me).
But then again that's my take on stuff :).
-Garrett
More information about the freebsd-questions
mailing list