Post by TPost by VanguardLHDoes the mobo support eSATA
one port
problem with eSata. Windows does not have the cool tools for mount,
unmount, and eject as does Linux. So, the customer has to plug
it in and leave it. eSata ain't portable under Windows. Sure
there are utilities for that, but they suck.
Wrong, wrong, and wrong. What were you thinking Obi-Wan ?
How can you have an ESATA connector on the back of a
computer, without Eject capability ? It was designed
for easy Plug/Unplug. The OS *has* to support that.
We can't have people running around ripping off
connectors, without some mechanism to do it safely.
Windows supports Hot Plug for ESATA and even for regular SATA.
In fact, it will even pretend you can eject C: but it
won't actually let you do that because (like Unix/Linux)
"the volume is busy" :-) These features were first
exposed when Hot Plug SATA first came out, and Microsoft
has had to clean up the UI a bit over the years, so as to
avoid scaring people. (People *hated* seeing SATA drives
offered for ejection... Especially seeing C: in the list.)
On any decent motherboard in a tower, look for an enable/disable
for Hot Plug in the BIOS. That can gate off availability.
You'll need to turn it on there first.
If you're serious about using a SATA dock for the customer (say),
there are articles on making this more convenient. You could
probably get a 5.25" tray that allows SATA tray operation, then
run a SATA cable from the back of that to the motherboard SATA
area. A superior tray for a tower, uses gentle lever mechanisms
for pulling the drive, as doing Hot Plug runs the risk
of mechanical damage if the spindle is still spinning.
https://superuser.com/questions/817226/sata-hot-plug-has-no-effect-in-windows
Use the small utility HotSwap!.
http://mt-naka.com/hotswap/index_enu.htm
"About HotSwap!
This software is developed based on the information Knowledge Base
Article #10744 - SATA: Hot Plugging Drives under Win2000/XP
(Internet Archive) provided by Silicon Image.
It does the same thing as you can remove device from Device Manager
but it provides much friendly user interface as you remove the
removable device from the "Safely Remove Hardware" icon in the
notification area. It also ensures that all data are written
and flushed to the disk before the device to be hot-swapped,
so you can use SATA/eSATA drive as a removable device much
alike USB/IEEE1394 drive.
"
That's the part I can't vouch for on Windows, is nicely
reproducible "parking" of hardware. For example, my
newest machine, if I "Safely Remove" a USB enclosure
with an SSD inside, the SSD keeps counting "dirty shutdowns"
implying for some reason it hasn't been put in a safe
state inside the enclosure. This is why it's generally
a good idea to buy enclosures with a USB LED that indicates
the quiescent state, before unplugging. Not all Windows
OSes seem to do this properly. (I can count on WinXP to
do it right!)
One of the hints in the SuperUser article, is to use Disk Management
to put a disk in the Offline state, to "encourage" a non-busy
hardware state. I've even had one report from someone who had
to do this for a *USB* drive - they can't safely remove a USB
HDD enclosure connection, without putting the drive "offline"
in Disk Management.
If you find a setup requires the "offline" trick, don't forget
that later when plugged back in, you have to set it "online"
again, because the system remembers the state across reboots.
My recommendation ? Play with it a bit more.
A "best" kind of design, would allow Safely Remove from
Windows, then, a power switch on the enclosure, so that
you can be absolutely certain the spindle is stopped
and the heads parked before moving it.
One of the reasons I don't like "toaster docks", is
because of the possibility of handling things while
they're still actually spinning. You want to test and
*make sure* (put ear to hard drive) that the software
is actually putting the drive in the desired state.
There's no point having a fancy "eject" setup, if you're
damaging the hardware doing it. Test test test... before
giving to a customer!
States:
1) Fully working.
2) Change drive to "Offline", which dismounts all partitions.
(There may be other ways to do a dismount besides that.)
3) Try safely remove, using some means.
4) Verify the damn spindle is stopping properly. Not
all hardware combinations seem to be handling this
properly (the proof being my SSD is reporting
"emergency power" events). You must *test* this is
working, before release to an actual customer.
If SMART reports more and more "emergency power"
type events, something isn't right, and you shouldn't
leave things in that miserable state. It could damage
something, if it was "for real".
Better than nothing, is a power switch you can use,
to switch off the power after (3), so at least any
rotating HDDs will be parked (even if it's
counted as an "emergency power" event). On consumer
SSDs, emergency power events counted are *not* a good thing.
Consumer SSDs don't have a supercap, and we don't
know how well they handle power events.
You could (eventually) brick the SSD drive.
5) Unplug "cold" drive and walk away with it.
You could do most of this testing on your home
setup (match the customer OS!!!), but I would
suggest verifying (4) onsite. Don't do a sloppy job.
I wouldn't have to offer warnings about (4), except
for the mixed results I'm seeing here. The results
don't inspire confidence, and could be due to the
enclosure chip design or firmware on my setup.
I did have SATA Hot Plug set up once here, but it's
not something I leave setup or enabled. I at
least proved it worked.
The first discovery of Hot Plug that I remember,
was by a USENET participant. He was relating how
he broke the SATA connector on a hard drive :-\
"And when he would hold up the SATA connector
against the hard drive stub, he could see his data
and get his files off." He transferred all the files
off the hard drive, by holding the connector against
it. That was the first case of accidentally
noticing that Silicon Image had put Hot Plug
support in their driver, before Windows did.
It was pretty funny at the time. At the time,
we wouldn't have predicted that would happen
(no one would have suggested just holding
the connector against it like that).
Paul
