Hot Swapping IDE Drives

Ecological issues around computing. This is an experimental forum.

Moderators: Ralf Hutter, Lawrence Lee

Post Reply
woodsman
Posts: 78
Joined: Tue May 29, 2007 5:22 pm

Hot Swapping IDE Drives

Post by woodsman » Fri Jul 13, 2007 3:51 pm

Do any SPCRers have experience with the various IDE converter solutions? Here are two examples:

ICY DOCK MB123PK-B 5.25" Hot Swap IDE drive to SATA Aluminum Internal Mobile Rack

ICY DOCK MB123UK-B Aluminum body with partial plastic IDE to USB internal enclosure

Several reviews I have read about IDE converters are a mixed bag of results. However, I would not be running these IDE drives full time --- only to perform backups, disconnecting, and safely storing.

External IDE converters require bulky power cubes. The trays tap into the internal power supply connectors. Plug-and-chug.

Currently for backups I use mobile hard drive trays to insert and remove several IDE hard drives. Been doing this for many years. To my knowledge the IDE interface never supported hot swapping. So yes, I need to reboot every time I perform a backup. Rebooting never bothered me, however, because with my old hardware I accepted that as status quo. Besides, these are home office boxes, not 24/7 servers.

With new hardware the picture changes. Between SATA and USB, hot swapping is now the status quo. But I have all of those perfectly usable IDE hard drives that are used only for backups. With such little use, they all probably have long lives remaining. So how to mate these two technologies?

Seems with a USB/SATA-to-IDE removable tray I could swing the trick. A simple lock-key enables or disables power to keep the drive safe, but the underlying OS never would see the device because IDE does not support hot swapping. With these mobile racks, seems the problem is solved. Insert the IDE drive into the tray, flip the power switch, and through the tray interface, the underlying OS actually sees a new USB or SATA device.

In such an environment I could hot swap my existing IDE drives that I use for backups and not spend additional funds on that aspect of my new box.

None of these useful backup IDE drives exceed ATA-100. Although a SATA back-end is faster than USB, USB probably is just as good because the older IDE drives can pop only about 40 MBps anyway at ATA-100, which is about the typical USB 2.0 sustained rate. Still, I probably would side with the SATA solution because all modern motherboards have several SATA connectors and I need only one.

Any SPCR opinions on this strategy? Not exactly a silent solution with many older IDE drives, but certainly a green solution in that older IDE drives are reused --- and more green remains in my wallet too.

P.S. With respect to the two example trays, ignore the silly noise makers called fans. With both of my current mobile IDE trays I disconnected those nuisances long ago. The noise makers are not needed for backup drives that are inserted for only 15 to 20 minutes. And my BIOS is configured to spin down my hard drives after 15 minutes.

scdr
Posts: 336
Joined: Thu Oct 21, 2004 4:49 pm
Location: Upper left hand corner, USA

Post by scdr » Fri Jul 13, 2007 8:41 pm

Since you already have some mobile hard drive trays -
an even greener solution might be to stick the existing trays and drives in an old existing machine and use that as a backup server over network.
(If 100 base T too slow, might be able to network them via firewire, if both have firewire ports.)

Still have to reboot to swap drives, but just the backup machine being rebooted. Bit of net cable less expensive than new bays, also resources already spent to build the machine - so keeping it in use longer also helps environment. (Since will only be on while you doing backup, the fact that it takes more energy while it is on is probably not as significant an issue.)

Just a thought.

woodsman
Posts: 78
Joined: Tue May 29, 2007 5:22 pm

Post by woodsman » Sun Jul 15, 2007 12:34 pm

Since you already have some mobile hard drive trays - an even greener solution might be to stick the existing trays and drives in an old existing machine and use that as a backup server over network.
That might be an option I should consider.

Currently I'm the sole user of these boxes. With new hardware and virtualization, and one of the IDE-to-SATA converter mobile trays, I would not need to reboot my new box. I would need to reboot my NT4 Workstation because that guest OS never would see the new device and never supported hot swapping either. But "rebooting" a virtual guest OS is not a hard reboot but a soft restart of the guest OS. As the kids today say: easy-peasy. :) Unlike Windows with its infamous file locks, backing up a GNU/Linux box in real time is not an issue. So my thinking right now is that I buy an IDE-to-SATA tray, pop in one of my IDE drives, hotplugging services see the new disk, and I run my backup script. I also concurrently restart NT4 and perform my backups there too.

I realize I could have GNU/Linux backup NT4 through the virtualization network layer, but those Windows file locks still prevent real time backups without special software that is granted those rights. My long standing solution was much easier. Because I had to reboot to insert the second IDE drive, I merely booted to my alternate NT4 C: partition and then backed up my primary C: drive from there. No file locks. :) Another problem with backing up through GNU/Linux is restoring from backups. In such a case, those Windows files would be on ext3 partitions, not NTFS. So even if my virtual NT4 OS crashed hard and I needed to restore from backup, I'd want to do that from native NTFS partitions. Yes, there is today better support for NTFS through FUSE and NTFS-3g, but as far as I know, they do not support the older NTFS-4 format.

Initially I do not want to recreate the wheel. My Windows backup scripts are all well tested through the years so why not just keep using them? Ditto for my shell scripts for my Slackware backups. I already have my backup disks formatted so maintaining my current process seems the least resistive route.

With that all said, performing the backups through a network layer from one of my older boxes does maintain some appeal, however, if only for the geek or "cool" element. I do understand your reasoning about not buying the IDE-to-SATA tray and just use what I already have with no extra expense. That would be "green." I would have to put some thought into how I would need to revise my backup scripts, however, because they never were designed with networking in mind. I will spend some time thinking about that option before buying my new hardware.

I have mixed thoughts about retiring my old boxes because they are perfectly usable. My old 486 with 16 MB of RAM also is perfectly usable and yet has sat on the shelf for several years. At one time I considered converting the 486 into a firewall, but the box is completely ISA. To my knowledge ISA never supported 100Mbps, only 10Mbps. That speed is sufficient for the ISP, but way too slow for my LAN, which is at 100 Mbps. I'd have to buy a 100 Mbps switch to serve the network.

I could use one of the current two boxes for a firewall, which already support 100 Mbps network cards, but my Linksys WRT54GL with DD-WRT already does that and at a trifle of the energy consumption of a PC. So unless I discover a way to revise all of my backup scripts to a network box, I suppose the current two boxes will see little action other than occasional experiments. And that is not at all a green approach. :(

Still, just chatting about these ideas gets the brain moving in the (eventual) correct direction!

Trekmeister
Posts: 83
Joined: Wed Mar 08, 2006 8:29 am
Location: Luleå, Sweden
Contact:

Post by Trekmeister » Sun Jul 15, 2007 4:49 pm

Just how bad is "hotswap:ing" for regular IDE? This is probably a quite bad idea, might even be a really really bad one, but I'm throwing it out there anyhow.

You are running linux on the machine right? How about unloading and reloading the IDE-controller module after inserting a harddrive? I have never tried it but it might detect the harddrive, don't you think? Of course this would probably seriously mess up things if your system disk(s) are on the same controller... Have never used virtualised windows so have no idea how it accesses drives, however something tells me that if it is not through the guest OS, you would probably still have the same problem... meh!



And I share your pain about old boxes... Have quite a few of around too! Well I'm sure they will come in handy some day... or something... :wink:

BTW I think there actually was a 100mbit/s ISA card, but seeing how the ISA-bus is 16 bit/33MHz IIRC you would have a hard time getting more than 1/4 of the bandwidth. Or maybe it was VESA, or EISA, or MCA, or...!

woodsman
Posts: 78
Joined: Tue May 29, 2007 5:22 pm

Post by woodsman » Sun Jul 15, 2007 10:09 pm

Just how bad is "hotswapping" for regular IDE? This is probably a quite bad idea, might even be a really really bad one, but I'm throwing it out there anyhow.
Interestingly, I found some discussions in an old Linux kernel mailing list where some of the sharp folks contemplated how to do this. Additionally, here is what I found in The Software-RAID HOWTO:

=================================
Hot-swapping IDE drives

Don't! IDE doesn't handle hot swapping at all. Sure, it may work for you, if your IDE driver is compiled as a module (only possible in the 2.2 series of the kernel), and you re-load it after you've replaced the drive. But you may just as well end up with a fried IDE controller, and you'll be looking at a lot more down-time than just the time it would have taken to replace the drive on a downed system.

The main problem, except for the electrical issues that can destroy your hardware, is that the IDE bus must be re-scanned after disks are swapped. While newer Linux kernels do support re-scan of an IDE bus (with the help of the hdparm utility), re-detecting partitions is still something that is lacking. If the new disk is 100% identical to the old one (wrt. geometry etc.), it may work, but really, you are walking the bleeding edge here.
=================================

My mobile trays are keylocked to prevent physical electrical damage. But I have seen Windows and the Linux kernel go bonkers if I power down that drive while still running the OS. The BIOS also receives weird signals from the IDE controller if I power down the mobile tray and soft reboot. The BIOS will almost always display the infamous Press F4 to continue message when trying to scan the missing IDE drive. I have to hard reboot with the reset button to clear the message, or shut down.
BTW I think there actually was a 100mbit/s ISA card, but seeing how the ISA-bus is 16 bit/33MHz IIRC you would have a hard time getting more than 1/4 of the bandwidth.
I think I recall the same as you, and now that you jogged my memory I seem also to remember that the NIC was terribly expensive. Low market demand and all that. But I think you are correct that the ISA bus itself was a bottleneck and limited actual throughput.

I once wondered whether I could run my old 486 with a floppy based firewall such as BrazilFW. No hard drive to waste juice, and remove all the unnecessary cards. But I realized that even then the box would burn more juice than a modern off-the-shelf router. And the throughput problem still exists too. If the box was a few years older with a PCI slot or two, the story might have changed. :(

bobkoure
Posts: 191
Joined: Sun Aug 11, 2002 3:26 pm

Post by bobkoure » Sun Aug 19, 2007 1:01 pm

But... once you're using the IcyDock stuff (or other IDE to SATA), you're not really using the IDE bus but SATA - so you're not hot-swapping IDE drives at all, as far as your system is concerned; you're swapping SATA.

Post Reply