Hey guys, I'm currently using hddsuperclone, and all these are for recovering (slightly really) important files + learning value for me.

My current destination drive seems to be slowing down my 20% completed cloning process because I had it connected via USB instead of SATA.
or maybe I also suspect my current destination drive is just by itself slow or maybe failing now. (It's inside an enclosure I cant seem to open easily which I cant connect via SATA)

Will it be a better option if I get a new destination drive and continue cloning the remaining 80% for me to be able
to connect it via SATA? Will the data still be intact even if the destination has changed?

Can I also switch modes if I think the cloning is too slow?
When do you decide to switch to a different mode?


Here are some pictures that may help as honestly I don't know how to interpret these data

https://imgur.com/a/ZF6AShy
^ SMART

https://imgur.com/a/BZqbZkZ
^ Clone Settings for 1st Pass, I increased cluster size and Skip Size to 16384 because the manual says so for Samsung drives.
^ I used the Analyze function but I forgot to select the destination as null, is there any chance that will mess up my drive analysis?

From the analyze results it looks like the source is the cause of any slowness (possible slow responding issue, but it ran out of time to perform a significant test so I can't be sure). Plus I see you are using the reverse option, which usually slows the reading down by about a factor of 10. Maybe there is something else I can't see about the destination that you can see that is causing it to be even slower.


You would either need to clone the first destination to the new destination before continuing, or start over.


Yes, you can switch modes. But unless you have the paid version, you are limited in your modes, and the default auto passthrough mode will pick ata passthrough if the drive is connected via SATA, which is the best free mode.


No, you can use the normal destination when performing the analyze function. The only reason you would use the null destination is if you did not yet have a destination, and were just performing the analyze to assess the drive, and were going to start over with a new progress log for the actual recovery.

It would help more if you included a screenshot of the actual recovery, or better yet attached the progress log.

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

Thank you very much for responding!! I have a few questions.


Wait, I don't get it. How was it incomplete?
also
Judging from the analysis, is it possible that the problem was only the firmware or pcb that is causing the slowing? not a weak head?


I forgot, here's the previous cloning process and also the values it was showing on-screen

The analysis in the free version is limited to 2 minutes, and 30 seconds for the variance test. Your drive is reading so slow that it could not complete enough reads to actually know what is wrong.

Unfortunately the log you provided also does not have enough information to be able to diagnose the issue. What you need to do is start a new progress log with the default settings (you can turn up the skip size like you did, but leave all the other settings alone), and let it run for 15-20 minutes and post that. And if you want to know if it is some sort of firmware issue, do another run with a second progress log file and also post that. And for your own understanding, look into using HDDSCViewer to examine the log files. If the pattern is the same or very similar, then it is a physical issue. If the pattern is completely different, then it is likely some sort of slow responding firmware issue.

But wait, you mentioned a 20% complete cloning process? Where is that log? That would likely be much better to see what was going on, assuming it is not messed up in some way.

Just so you understand, the drive has some serious issues, and the more you mess with it (especially without knowing what you are doing), the more likely it is to die. If the data is important enough to pay for, you should seek professional recovery. If you continue with DIY recovery, then you are accepting the risk of loosing the data.

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

SAMSUNG HN-M320MBB only has 2 heads chances are all the test was done on head 0, even backwards those read times would be sky high. Set the skip size for 20GB or so (50mill) and do a phase 1 pass with everything else as default. Post the log so we can see if and where the drive is reading or not.

BUT as Maximus has already said.
If the data is important enough to pay for, you should seek professional recovery. If you continue with DIY recovery, then you are accepting the risk of loosing the data.

_________________
Properly, Quickly, Cheaply - Pick any two.

@Lardman, thanks for taking an interest in this. I think you also understand that we don't have enough data to know what is happening. There are a few things that indicate it could be a weird firmware issue, but that is not conclusive in any way. And since Samsung drives are not well know for a "slow responding issue", this could very well be a bad/weak head issue. I don't always have the time to assist with cases like this, so your input is very welcome

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

I appear to have nothing but time at the moment, might as well make use of it As you know I put hundreds of test disks through HDDSC, I then put the same disks through PC3K to compare the results, behaviour and diagnosis. I'm happy to help field the ground balls or to butt out

_________________
Properly, Quickly, Cheaply - Pick any two.

Hello, sorry for mentioning that, the log file was actually the real result because I thought it sounds ridiculous or embarrassing if I said I just wanted to switch to a new destination drive when the cloning process was still 0.001495%, I don't know if it was possible. I guess I've come to a conclusion that my drive has a weak head than a firmware problem since I hear clicks but they all disappear and comes back again. If it isn't too much to ask to wonder, why/how do weak heads damage a hard drive? I don't see it being mentioned everywhere and it is quite counterintuitive because the term "slow" just sounds less destructive?

Stuff like clicking/scratching/ticking or beeping are essential bits of info for making a diagnosis.

There's a world of difference between a weak head and a damage one. A quick google should show you the physical damage a head can do to a platter, you can't get data back from dust. A slow responding drive isn't normally just "a slow problem" itself it's a result of other problems causing the drive to be slow.

_________________
Properly, Quickly, Cheaply - Pick any two.

With the new information, I would like to revise my advice. First, if the data is important enough to pay for, stop now and don’t listen to anything else I say, except to seek professional recovery. Since you did say the data is only slightly important, and indicated you want to learn, then I will provide further advise that could cause permanent data loss.

Start a new recovery (new progress log), with the destination that you really want. Disable all other phases except 1-2, to get the most good data from the (hopefully) good head. Since it is a Samsung drive, and given the log provided so far, I think it will be okay to change the skip size to 16384. Leave all other settings default. Start it and let it run. Maybe post the log after it runs for an hour so we can see how it is going. Watch for skip runs, which will indicate the bad/weak head. You should be able to see a pattern using hddscviewer. If the other head reads normal, you should end up with about 50% recovery in a few hours or so, maybe longer depending on how the drive is responding. As long as the settings are good enough and the drive is in the expected condition, the self learning head skipping algorithm should do its job to get to the 50% mark. After that, I am not sure that you will have much more success, assuming you even get that far without the drive dying.

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

@Maximus Thanks, I have a question before I retry cloning.

Will increasing the cluster size also help? Or is it just useful for specific drives?

Should I switch my BIOS setting to IDE rather than AHCI now that I will use the Passthrough ATA mode?

I also found the screen output from the manual can be useful in my next clone.
Will this be the right command for logging all screen output to a file? Assuming I'll save it to a flash drive.

script -c "sudo hddsuperclone" /media/xubuntu/TROJAN/linux/panibago/hddsuperclone.log ?

At this time for your case there does not appear to be any benefit to increasing the cluster size. Leave it default for the phase 1-2 attempt.

No, leave the BIOS in AHCI mode. There are only specific reasons for IDE mode, and it will not benefit you at all.


I think so, although I don't use the command often enough to be familiar with it. I would like to point out that with that command, if you had to restart hddsuperclone I think it would overwrite the last saved terminal log. I think it would be safer to add the -a option so it would append the new lines, otherwise you would need to remember to use a different log name each time.
script -a -c "sudo hddsuperclone" /media/xubuntu/TROJAN/linux/panibago/hddsuperclone.log

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

Since you are interested in advanced logging, I would also enable the phase logs in the advanced settings.

And I would like to add that if it stops because of a skip reset, then post the log before continuing so we can analyze it.

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

Wait, I forgot to add. I suddenly remembered these.
Maybe these will provide some useful information.

This was the smart CrystalDiskInfo from Windows of the Samsung Source drive back when I haven't heard of hddsuperclone yet.

The SMART status just indicates that the drive is failing. But you already know that. There is nothing else useful about it, as it is not capable of providing any info on what the failure actually is.

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

Thank you very much and for answering my specific questions,
I tried cloning only one partition of this source drive from ddrescue months ago, and was ready to give up on it for only rescuing 25% data, but I started from scratch and got 46% this time with hddsuperclone (for now).

Is this because ddrescue lack the functionality to skip bad sectors than hddsuperclone?
I guess one thing that separates this from ddrescue is its Skip Resets.


Here's what Ive done so far after so many hours of reading forum posts

Screen display of 1st run (2 hrs in)


so....after a couple of hours cloning, a power outage shut my computer off and I'm not quite sure to what position it stopped after the sudden shutdown. But it seems the drives are still functional after the incident

I then tried making a backup of the ".bak" file of the original "Feb14-image.log" log file to my Windows pc in case something happens again.
But then chkdsk kinda destroyed the original bak file without me knowing.
Thank God, there is still a bak file that remained other than the main one,
which is "Feb14-image.log.Phase 1.2021-02-14_11.13.08.617101.bak" and I used that for the next re-run.



Here appears the new weird progress log filenames after repairing them via hddsc.
This is where I got confused, how do you differentiate which is which?




For the 2nd run (the run after power outage), I picked the only remaining ".bak" log file
"Feb14-image.log.Phase 1.2021-02-14_11.13.08.617101.bak" because it seems the most recent among the four.
What's the difference of this with the ".617101" log file anyway?




Screen display, 14 mins after the power outage (2nd run)


Screen display, 5 hours after the power outage (2nd run)


The skip size has also increased even though I didn't change it, how did this happen?


The progress log file ".bak" file (I renamed it to ".log" because .bak isn't allowed here)


Why is the "size on disk" smaller than the actual file size, is this normal?


I stopped doing the whole process because it's current rate now slowly transfers at only 50-600 KB/s.
Is there any way to speed up this process?

TL;DR

Head - looks weak rather then completely dead though, you may get lucky if you just wait it out. Untick the scrape/divide/trim, Id bump the skip threshold to 5000 and fully complete phase #2 if you haven't yet. Phase 3 even with the default settings will show if the head is dead or not if none of the black area gets filled in the first 20%. If you start getting data in the skipped areas we can tweak the settings to balance performance / recovery, 50-600K is far better than nothing and a lot cheaper than a head swap.

_________________
Properly, Quickly, Cheaply - Pick any two.

If you look closely, the phase log file name tells you what phase it was at the end of (or currently in if the program was stopped), and the rest is a date-time stamp, so you can tell exactly what time it was written (down to the microsecond). And if you had to use one for recovery because the original was lost, you should have made a copy of it and then renamed the copy to the original name. You are making a mess of file names that is hard to understand. And when you lost the .bak file, what about the original file “Feb14-image.log”, did you loose that too?

The skipping algorithm changed the skip size. The original size is still remembered in the settings in the log, as the min skip size.

The size on disk is smaller because of how the writes are done. The file is “sparse”, meaning the 52% that has not been written. If you were to read an area of the file that has not been written, it would read as zeros. If you make a copy of the file, the copy will inflate to full size.

You need to uncheck Trim, Divide, and Scrape. When I said to only do phases 1-2, I meant uncheck ALL of the other phases. If it has got into trimming, you need to stop. If not, then uncheck everything except phase 1-2 and continue until phase 2 completes. Phase 2 is a compliment to phase 1 to get the most data out of the good head.

More in another post after I look at the log.

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

I don't agree with bumping up the skip threshold, that will stop skipping and defeat getting the most good data first, and make the rest of phase 2 take much longer.

Once phase 2 is complete, then you should uncheck phases 1-2 and check phases 3-4. It may then also be time to test whether or not increasing the clustersize helps or not. Run it for a few minutes and watch the speed and read times, then jump the clustersize up. Try doubling it at first, and if it helps keep doubling it until it either doesn't help or you hit the maximum. You want the read speed to increase significantly, but the read times should not go up much. If it doesn't help (or makes it worse), then you probably should turn it back down.

If it does help, you may get a decent read speed. The very beginning of the drive is the worst, but after about the 7GB mark it gets better and seems fairly consistent. Keep that in mind when adjusting the cluster size. Once the position is past 7GB you should be able to see any difference when adjusting.

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool

When changing the clone settings (unchecking and checking things), a side effect will be that the min skip size will be overwritten with the current skip size. So before you start phase 3, you will need to change the skip size back to original.

I think I need to figure out how to fix that so it only changes the min skip size when the skip size is changed by the user, instead of when any of the clone settings are updated.

_________________
http://www.sdcomputingservice.com
Home of HDDSuperClone and HDDSuperTool