New Linux Tool - HDDSuperTool

[maximus]

1. Better documentation on what the -C / --command flag actually does and maybe a couple real-world examples. The docs here are too sparse.

I will see what I can do for that.

2. Better parsing/understanding of filenames passed via the -f / --file command.

Will be fixed in next release (code is implemented and tested). It will extract the directory path from the script file name that was entered on the command line, and will look for all includes in that same path.

3. Proper use of exit codes.

Oops! It is supposed to exit with something else other than 0 if there is an issue. Only a few lines of code to fix that oversight, it will be fixed in the next release. I think I broke that when I implemented the exit code capability from within the scripting. With that being said, I do not have my exit codes documented or properly set up to be differentiated. For instance failing to open the device will result in an exit code of 255 (-1). More work to be done on that in the future.

On your site you provide some "helpful tips" that reference using tee and things like 2&>1. Might I suggest mentioning "script"

Nice! I was not aware of that command. I think it is likely much better than tee although I will have to test it more. And you don't have to remember to exit if you use the --command option.

I do not know when I will have a new public release ready. I have been focusing on the direct IDE and AHIC modes and getting a private beta version ready for a limited number of testers. I have made so many code changes that I will need to put it through all sorts of tests before I can make any sort of public release.

 

[maximus]

It is finally time for beta testing! The link to get more information is below. Available files will be password protected, and you will need a valid license key. At this time the only way I am providing that information is through PM on this forum. Please request such information from me if you are interested and capable of the beta testing.

Please note that this testing is using the scripting ability of HDDSuperTool, and is not yet any sort of actual cloning tool. But this testing is for the backbone of the concept, so it will directly impact any move forward on a cloning tool.

http://www.sdcomputingservice.com/hddsu ... ta-testing
Note that the page that this link points to may be deleted, or the permissions turned off, at some time in the future if I feel it doesn't need to be there anymore. If I make any changes on the site that need to be mentioned I will do so here.

 

[maximus]

1. Better documentation on what the -C / --command flag actually does and maybe a couple real-world examples. The docs here are too sparse.

I am adding a simple example to perform an identify device command. But I obviously have not even tried the -- command option in a while as I found it to be broken. It was not opening the disk for access in passthrough mode so it would just give an IO error when trying to send commands. I am not sure when I broke it, I know it used to work :? It will be fixed in the next version.

 

[Jared]

Given the popularity of this thread, I'm going to create a separate section for this tool on the forum. That way you can start sub topic discussions about the tool and it's features if you want instead of having one thread that's 30 pages long.

 

[maximus]

Just an update on the cloning tool. I have a working alpha version (command line only) that seems to be working and has passed an initial data integrity test, although it is not ready for the public yet. I am thinking there will be three separate versions:

A free version that is command line only and only has passthrough access, maybe only the scsi level. It will be possible to import a ddrescue log (map) file, and even export also in case someone does not like my cloning tool for some reason and wishes to switch back to ddrescue. This could be released first (maybe in a few weeks) to get people interested, as it should be better than stock ddrescue (will be able to tell when a drive drops out due to device fault and exit with a specific code and message so the user can power cycle the device and resume).

An affordable paid lite version that will only have passthrough access, but will have the GUI and be able to directly control a certain relay for power cycles. There will be an option to purchase a relay setup directly for extra cost. I will also provide instructions on what relay will work and how to hook it up if someone wishes to make their own setup. When I start implementing the ability to recover files and folders (at least with NTFS) it could also be included with this option for an extra cost.

Then there will be the pro version that will come with relay hardware, modified IDE cable for hard resets, and will require a USB dongle to run (license management). This is the one that can access the disk directly either IDE or AHCI, and control the reset timers for the different phases of recovery

Much work yet to do, but that is the plan so far.

 

[pclab]

Very good roadmap.
Keep it up.

 

[Jared]

Brilliant! Just Brilliant! I think this sounds like a plan which could really work.

 

[maximus]

Here are a couple screenshots of ddrescueview analysis of a smart self learning head skipping algorithm I am working on. This drive has only one platter with 2 heads, and actually only has a weak/damaged head that has many single sector errors, but this drive is actually 99.9% recoverable. The normal algorithm will try to settle in to only doing about 7 reads in the bad head until it skips to good data, and the aggressive one will do about 4 reads in the bad head (this number is per pass). This is done with one forward pass and one reverse pass with skipping. It is still not the same as only reading the LBAs from good heads (aka turning a head off), but it is about the best one could hope for without that ability. It is designed to get the most good data first and as fast as possible, but without missing too much of the good data. There is still more testing to do, but I was pretty happy with this result.

 

[abedalkareem]

[post]4304[/post] This is done with one forward pass and one reverse pass with skipping.

very nice feature , its useful to develop DR head skipping algorithm reliable enough to adjust it self according the hdd condition and the location of bad sectors , especially when the HDD start imaging normal and stable and left it hole week end imaging and comeback to find it unstable and hits bad area while the good head ripped off

[glow=silver]great work keep it up[/glow]

 

[maximus]

Even so, it is still flawed. If you look at both, there are a few random areas in the good head that have errors and had some skipping happen. This condition can easily happen if a timer is set low to speed things up, and there are any somewhat slow areas (likely what happened in my results). Now to figure out how to mark the likely bad head areas so any other areas in the good head(s) can still be further processed. Already have an idea to try... :twisted: