On Wed, 01 Jun 2005 10:16:29 -0500, Thorn <Raptor@Etruscan.Warrior> wrote:
>Hdparm is not yet available for 2.6 kernels.
I must have imagined this:
login as: root
Authenticating with public key "grant@peetoo" from agent
Last login: Thu Jun 2 04:39:48 2005 from magpie.mire.mine.nu
Linux 2.6.11.11a.

I just know I'm a better manager when I have Joe DiMaggio in center field.
-- Casey Stengel

peetoo:~$ hdparm -i /dev/hda

/dev/hda:

Model=ST3120022A, FwRev=3.54, SerialNo=5JS1SGLT
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs RotSpdTol>.5% }
RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4
BuffType=unknown, BuffSize=2048kB, MaxMultSect=16, MultSect=off
CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=234441648
IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120}
PIO modes: pio0 pio1 pio2 pio3 pio4
DMA modes: mdma0 mdma1 mdma2
UDMA modes: udma0 udma1 *udma2 udma3 udma4 udma5
AdvancedPM=no WriteCache=enabled
Drive conforms to: ATA/ATAPI-6 T13 1410D revision 2:

* signifies the current active mode

peetoo:~$

--Grant.

On Wed, 1 Jun 2005 17:38:27 +0300, "Mr.Jason" <jaakkochanREMOVETHIS@surfeu.fi> wrote:
>hdparm -y /dev/hdb

Big Y for shutdown drive, prior to that though, unmount all
partitions on drive, tell drive to spindown, pull data lead,
pull power lead. Nothing will fry, no oops 'cos filesystem not
mounted. If you talking RAID1, I have no idea.

Another drive on same ribbon might get confused --> dataloss.

Sort of reverse order going other way, power on, data connect,
not something I would expect to work, need BIOS support, AFAIK
linux still needs reboot to re-read drive geometry, unsure.

I'd also suggest USB enclosure, safer, saner, and cheap

--Grant.

Grant Coady wrote:

> ...
> not something I would expect to work, need BIOS support, AFAIK
> linux still needs reboot to re-read drive geometry, unsure.

Using hot-swappable SCSI drives, I've used "sfdisk -R /dev/$disk" to get
the kernel to read the partition table of a newly inserted disk. That
has worked flawlessly for me.

--
----------------------------------------------------------------------
Sylvain Robitaille syl@alcor.concordia.ca

Systems analyst Concordia University
Instructional & Information Technology Montreal, Quebec, Canada
----------------------------------------------------------------------

On Thu, 2 Jun 2005 05:06:21 +0000 (UTC), Sylvain Robitaille <syl@alcor.concordia.ca> wrote:
>
>Using hot-swappable SCSI drives, I've used "sfdisk -R /dev/$disk" to get
>the kernel to read the partition table of a newly inserted disk. That
>has worked flawlessly for me.

Out of my league I got bitten once with couple IDE drives, didn't
reboot after playing with partition tables prior to formatting some
partitions, wrote 200 GB (80 + 120) zeroes to drives and started over.
It was spending much of the following week loading about 50GB from
backup CDROMs that made the experience stay in mind.

--Grant.

Mr.Jason wrote:

> Would it make difference if the disk is on mobilerack or
> not?
> Professional electricians are welcome to join this conversation and give
> your opinions!
>
> Mr.Jason

[recent story]
As I do some computer adjustments, I don't call this repairs, I see
people who "will try anything".
So a "deflated" looking gentelman appears holding a plastic bag with HD
in it.
A brand new 40Gb, which he "tried" to hot-swap, by his story it gave a
puff of smoke and went on strike.
OK, we found the proper tools to remove the special screws that hold the
printed circuit in place, carefully (I do this VERY rarely) lifted the
board and discovered a blown surface mounted IC. ( Of the 1cm/sq and
about 80 connecting points kind).
Being resourcefull we ordered another HD, exactly the same model so we
can swap the circuit and recover the treasures hidden on plates.
When it arrived, we swapped the electronics, NOT hot-swapped it to a
working box, the BIOS recognised it but the sounds coming from the
mechanics were of the kind you hear in horror movies.
[/recent story]

"Your mileage may vary." as they say.

Have fun

Stanislaw
Slack user from Ulladulla.

On Thu, 02 Jun 2005 09:59:38 GMT, Stanislaw Flatto
<compaid@shoalhaven.net.au> wrote:

> A brand new 40Gb, which he "tried" to hot-swap, by his story it gave a
> puff of smoke and went on strike.

Gruesome.

> OK, we found the proper tools to remove the special screws that hold the
> printed circuit in place, carefully (I do this VERY rarely) lifted the
> board and discovered a blown surface mounted IC. ( Of the 1cm/sq and
> about 80 connecting points kind).

That seems to support 'latch-up' someone else reported, was it a nice
crater in chip, or just a gentle crack that let the magic smoke out?

They don't make chips like they used to[1]: Years ago I was called down
to factory floor -- i8748 EPROM uC not working... Turn it on and the
corner chip die power connects glowing red under chip window...

Chip in backwards, turn off power, put chip in right way around and
it worked. (reverse connect power to chip forward biases the substrate
parasitic diode), this was in mid to late eighties, soft power supply.

[1] Now I'm starting to sound like an old fart.

--Grant.

Grant Coady wrote:

> They don't make chips like they used to[1]: Years ago I was called down
> to factory floor -- i8748 EPROM uC not working... Turn it on and the
> corner chip die power connects glowing red under chip window...
Those were VERY good, at least you could "see" what you wrote in...
>
> Chip in backwards, turn off power, put chip in right way around and
> it worked. (reverse connect power to chip forward biases the substrate
> parasitic diode), this was in mid to late eighties, soft power supply.
This is called "reverse engineering"
>
> [1] Now I'm starting to sound like an old fart.
Maybe we should start a club.
>
> --Grant.
>
Stanislaw

On Wed, 01 Jun 2005 22:17:12 GMT, Grant Coady Cried: Read These Runes!:
> On Wed, 01 Jun 2005 10:16:29 -0500, Thorn <Raptor@Etruscan.Warrior> wrote:
>>Hdparm is not yet available for 2.6 kernels.
> I must have imagined this:
> login as: root
....

Probably not. Hdparm is kernel irrelevant, but the specific function I was
referring to (hdparm -b0 /dev/xxx) has not yet been implemented for 2.6
kernels. I recall reading Alan Cox saying he might get around to it this
month (June) time permitting.

A sin of omission? Thanks for prompting me to clarify.

Thorn
--
Every four seconds a woman has a baby. Our problem is to find this
woman and stop her.

On Thu, 02 Jun 2005 09:44:41 -0500, Thorn <Raptor@Etruscan.Warrior>
wrote:
> > I must have imagined this:
> > login as: root
> ...
>
> Probably not. Hdparm is kernel irrelevant, but the specific function I was
> referring to (hdparm -b0 /dev/xxx) has not yet been implemented for 2.6
> kernels. I recall reading Alan Cox saying he might get around to it this
> month (June) time permitting.
He's one busy person You just triggered a kneejerk reaction from me.

I wish hdparm could see further into IDE drive mounted in USB enclosure
and/or somebody write drive shutdown accessor for it. I really would
like to spindown HDD before disconnect. And the idea of ripping it out
of USB enclosure, mounting in a box with adaptor just to change drive
spindown timer strikes me as a wierd way to fix a software problem,
unless it be the USB bridge chip does not convey full command set...
>
> A sin of omission? Thanks for prompting me to clarify.

Nah, just for fun ) 'hdparm -b' -> get/set bus state? me puzzled?

--Grant.

On Fri, 03 Jun 2005 04:30:20 +1000, Grant Coady Cried: Read These Runes!:

> Nah, just for fun ) 'hdparm -b' -> get/set bus state? me puzzled?

Using a removable device tray for IDE devices it's recommended to turn the
bus off before powering off the tray and removing the device.

Example:

hdparm -b1 /dev/hdf

/dev/hdf:
setting bus state to 1 (on)
busstate = 1 (on)

And dmesg shows:

Probing IDE interface ide2...
hde: CREATIVE CD-RW RW1210E, ATAPI CD/DVD-ROM drive
hdf: WDC WD400BB-00CAA0, ATA DISK drive
ide: drives found on hot-added interface.
hde: attached ide-default driver.
hdf: attached ide-disk driver.
hdf: host protected area => 1
hdf: 78165360 sectors (40021 MB) w/2048KiB Cache, CHS=77545/16/63, UDMA(100

This process is part of a script I use for rsync'ing my working installation
to a removable drive.

Thorn
--
I brake for chezlogs!