I have lately been reading about "Safely Remove Hardware" option in Windows. Here are some details I have found about this feature -
(1) Was first introduced in Windows ME. Has since been a part of Windows 2000 and Windows XP. Windows 98, unfortunately, lacked this feature.
(2) As Garycase has clarified on Experts Exchange -
There are two issues associated with removing a USB device: (1) possible data corruption; and (2) possible electrical damage. For (1) the problem is that Windows does deferred writes to removeable devices, and if you remove one before it has completed the write process (which can be several seconds after it's "apparently" done) you will have a corrupted drive. As noted above, if you change the settings for the device to optimize it for quick removal, you can disable the write caching feature and then you don't have to worry about this ==> but it will make writes MUCH slower, particularly when copying many small files, as it will close the directory after each file. For (2) there's no workaround -- you're simply taking a risk every time you pull the device out; and any electrical "glitch" can damage either the device itself or your PC's USB port. There are MANY examples in questions here on EE where folks have destroyed USB flash drives; their USB ports; an MP3 player; etc. by not following the "Safely Remove" protocol.
Bottom line: Think of "Safely Remove" as a seat belt. You'll "get by" without using it MOST of the time; but at some point you'll have an accident (i.e. you'll damage your device) and then you'll wish you had used it. If you aren't patient enough to do so, just keep "yanking away" ==> but don't be surprised when your device fails.
Additional note: If a device is self-powered (i.e. uses no power from the USB bus), and is ONLY used for reading data from the device (no writes), then it is much safer to simply "pull" the device. Common devices in this category are digital cameras, or compact flash cards used in powered external readers (e.g. a card reader slot on an all-in-one printer). In those cases, it's "okay" to just pull the device ==> since there's no power issue with the bus and no deferred writes (since nothing's been written to the device).
(3) The "data corruption" issues asssociated with not using "Safely Remove Hardware" option are also discussed in detail over here -
Changes in Windows XP Beta 2
Refinements to Caching Policy for Consumer Storage
Essentially, anything a user can physically remove from a machine could be removed from a running system without prior warning to the operating system - for example, any kind of device on CardBus, PC Card, USB, IEEE 1394, and so on. However, the area of greatest concern is storage devices, because of the possibility of data loss or corruption when surprise removal of these devices occurs.
To mitigate the likelihood of data loss in these scenarios, Windows XP has a refined caching policy for removable storage. As of Windows XP Beta 2, for consumer-oriented removable storage (USB, Flash, Zip, and so on), write caching is disabled by default.
Disabling write caching means that, instead of saving up changes for a file on a removable storage device and then doing a bulk write, Windows XP writes changes to the file as the changes are made. This keeps data on removable storage devices more current, mitigating the likelihood of data loss. However, disabling write caching also has a performance impact. Therefore, for IEEE 1394 hard disk drives, Windows applies a special case for the default caching policy by not disabling write caching for these devices.
The basic way to think of write caching policy is:
- If write caching is disabled by default, the storage device in question is optimized for safe surprise removal. This is the appropriate action for slower, consumer-oriented storage devices such as flash-type memory.
- If write caching is enabled by default, the storage device in question is optimized for performance. This is the appropriate action for high-performance storage such as SCSI and IEEE 1394 hard disk drives.
Note that caching policy defaults for a given device can be changed in Device Manager.
Optimizing the Default Caching Policy
As noted earlier, in Windows XP Beta 2, IEEE 1394 hard disk drives have write caching enabled by default, to optimize for performance rather than for safe surprise removal. This also means users should first stop IEEE 1394 hard disk drives using the hotplug applet before removing them. However, the Beta 2 mechanism for special-casing IEEE 1394 hard disk drives requires additional refinement.
This content has been culled from the Microsoft site - http://www.microsoft.com/whdc/system/pnppwr/hotadd/XPrem-devs.mspx
(4) Is applicable for all hot swappable devices like - IEEE 1394 devices and USB devices.
(5) In case, you got a hot swappable mass storage device plugged in and you do NOT get the "Safely Remove Hardware" option to STOP the device before unplugging it, you can use the following instructions to manually invoke the "Safely Remove Hardware" option -
You can click on Start, Run..., and type in:
RunDll32.exe shell32.dll,Control_RunDLL hotplug.dll
And press OK.
That will bring up the "Safely Remove Hardware" dialog.
A side effect could be that the "Safely Remove Hardware" icon reappears as it should.
If it happens often enough, you can create a desktop shortcut for this geeky looking command.
These instructions have been pulled from the site -
http://ask-leo.com/safely_remove_hardware_where_did_the_icon_go_how_do_i_safely_remove_hardware_without_it.html
(6) Bottom line -
If it is a hot swappable mass storage device, you MUST stop the device before you unplug it. In this case, you might go to the extent of manually invoking the "Safely Remove Hardware" wizard.
If it is just any other hot swappable device and is self powered, go ahead, unplug it without any need to STOP the device.
If it is a hot swappable device, NOT self powered and NOT a mass storage device, see if you can locate the "Safely Remove Hardware" wizard. If you can, stop the device before you unplug it. If you cannot, well go ahead and yank it out anyway.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment