conserving information from a falling short drive
An exterior 3 1/2" HDD appears to be at risk of falling short-- it is making ticking sounds when idle.
I've gotten a substitute drive, and also need to know the most effective approach to get the information off of the uncertain drive with the most effective opportunity of conserving as high as feasible.
There are some directory sites that are more vital than others. Nonetheless, I'm presuming that deciding on directory sites is mosting likely to lower my opportunities of conserving the entire point. I would certainly additionally need to place it, unload a documents listing, and afterwards unmount it in order to have the ability to properly focus on directory sites. Including the reality that it is time - consuming to do this, I'm leaning far from this strategy.
I've taken into consideration simply making use of
dd, yet I'm not exactly sure just how it would certainly take care of read mistakes or various other troubles that could protect against just particular components of the information from being saved, or which can be gotten over with some retries, yet not numerous that they jeopardize various other components of the drive from being conserved. I presume preferably it would certainly do a solitary pass to get as high as feasible and afterwards return to retry anything that was missed out on as a result of mistakes.
Is it feasible that replicating extra gradually-- as an example stopping every x MB/GB-- would certainly be far better than simply running the procedure full throttle, as an example to stay clear of any kind of getting too hot concerns?
For the "where is your back-up" group : this in fact is my back-up drive, yet it additionally has some non - essential and also cumbersome things, like songs, that aren't back-ups, i.e. aren't supported.
The drive has actually not shown any kind of clear indicators of failing apart from this rather threatening audio. I did need to fsck a couple of mistakes lately-- orphaned inodes, wrong free blocks/inodes matters, inode bitmap distinctions, absolutely no dtime on removed inodes ; concerning 20 mistakes in all.
The filesystem of the dividing is ext3.
There is no other way of recognizing the most effective of your alternatives without recognizing specifically what is failing with the drive. If it is a mechanical failing, staying clear of home heating it up can aid, yet if it results from mistakes in the servo information, warmth isn't most likely to matter.
I would quickly start replicating the one-of-a-kind information to the new drive with
rsync will certainly allow you stop, resume, and also reactivate as essential till you get all the information off.
After that I would certainly run an information scrub on the drive. I think from the
ext3 filesystem that you are running Linux, so attempt this:
# umount /dev/sdX # badblocks -n /dev/sdX
(Unmounting the drive first is necessary.)
This will certainly read every field from the disk and also write it back without adjustment. That will certainly compel the drive firmware to examine every field for mistakes and also to remap any kind of negative fields it locates. This is one of the most integral part of what the pricey SpinRite program does. Tip up to that just if
badblocks falls short and also you still have not obtained all the one-of-a-kind information off the drive : SpinRite attempts tougher than
You can make use of
myrescue to duplicate the falling short disk, without terminating on any kind of unreadable field. (Myrescue is much less configurable yet has a far better default approach as it attempts to miss over unreadable areas.) This will certainly replicate every little thing consisting of empty room and also will not allow you set top priorities. Nonetheless, such a reduced - degree strategy has a benefit over filesystem - degree devices: if a directory site is unreadable, you could still recoup the documents it has by looking the raw photo with devices such as
photorec consisted of in testdisk, etc