1 Topic by ronbaby

  • ronbaby
  • Member
  • Offline
  • Registered: 2012-09-04
  • Posts: 49

Topic: Expected behavior? (pATA "hot swap" racks)

Assume the following scenario...

Assume you have an ordinary sort of (older) motherboard which has built-in Parallel ATA (pATA) connectors on the motherboard.  Assume also that you own one of these "hot swap" pATA racks:

http://www.amazon.com/Genica-Mobile-Rac … B000FA0W80

(Note that on the back of the boxes these things come in, they explicitly say "Truely hot swappable - you can insert and remove with your computer power on.")

Assume further than you have correctly wired up one of these things to your motherboard, and that you have two (2) of the swappable carriers, each with a different pATA hard drive contained therein.

Assume that one of these is inserted (and locked in, i.e. power enabled) into a powered down system, and that the system is then powered up and caused to boot from a recent vintage GParted Live CD (whereupon GParted will perform its normal startup device scan and will correctly find and identify the plugged in drive).

Now assume that one turns the unock key (which powers down the currently inserted drive) and that one then removes the current carrier+drive and then inserts the #2 carrier+drive and that one then again turns the key to the "ON" position.

Finally, assume that after all this, one clicks on the "Rescan devices" button within GParted.

What is the exepected behavior of GParted Live CD in this scenario?

Would completely locking up, i.e. such that no more activity (on the screen) is visible, and such that no more keyboard and/or mouse input has any apparent effects any longer be considered within the range of expected behaviors of GParted under such a scenario?

2 Reply by gedakc

  • gedakc
  • Developer
  • Offline
  • Registered: 2007-11-30
  • Posts: 4,593

Re: Expected behavior? (pATA "hot swap" racks)

If USB drives are inserted or removed, the Linux kernel recognizes the device change and a subsequent GParted rescan should list the device.  As for PATA drives, I don't know what would happen.  It would not suprise me if the entire GNU/Linux system locked up.  As far as I know, PATA drives were not originally intented as hot-swap when first added to PCs.

3 Reply by ronbaby

  • ronbaby
  • Member
  • Offline
  • Registered: 2012-09-04
  • Posts: 49

Re: Expected behavior? (pATA "hot swap" racks)

gedakc wrote:

As for PATA drives, I don't know what would happen.

Not to be flippant, but I think that I just told you what happens.

It would not suprise me if the entire GNU/Linux system locked up.

Bingo!

So to whom should I report this rather entirely unfortunate bit of behavior?  Linus?

As far as I know, PATA drives were not originally intented as hot-swap when first added to PCs.

Undeniably true.  But what has that got to do with anything?

There is nothing in-built into the hardware that is causing the total lockup.  It is a strictly software non-feature, and not, as far as I can tell, an entirely unavoidable one.

4 Reply by gedakc

  • gedakc
  • Developer
  • Offline
  • Registered: 2007-11-30
  • Posts: 4,593

Re: Expected behavior? (pATA "hot swap" racks)

When you swap the drives, does the Linux kernel recognize the change?  You might check the output of dmesg for evidence.

5 Reply by ronbaby

  • ronbaby
  • Member
  • Offline
  • Registered: 2012-09-04
  • Posts: 49

Re: Expected behavior? (pATA "hot swap" racks)

gedakc wrote:

When you swap the drives, does the Linux kernel recognize the change?  You might check the output of dmesg for evidence.

I would have enjoyed checking that, but alas, it is too late now.  Yesterday (Sunday) and today (Monday) I tore apart the old machine that had pATA ports and upgraded the guts to a new motherboard, new CPU and new memory.  The new motherboard has no pATA ports at all.

Oh well.