Need assistance with comming back online asap

I share your opinion that the description could be more precise. I got lost by myself and had to correct my previous post. But it was not an easy situation, so I can understand @Ruskiem, all those operating systems, interfaces and even a new controller. Great point about SystemRescue and particularly about Windows 7, but even if it was made for up to 2.1T on GPT so AFAIU this disk should work assuming BIOS is supporting it. Anyway, I am a bit sceptical but at the same time if there is an effort going I hope for the best.

Was it win7 before you trying an another controller?
If it was win10, then it uses extended NTFS features like dedupe and compression by default, and they are not supported well under win7 (except compression), they also doesn’t supported out of the box in the ntfs-3g Linux package (you will have to compile it with special options and configure plugins), so any Linux restoration tools for NTFS likely will corrupt it even more.
If that’s the case, you need to connect it back to a previous PC and use the same controller and try to fix it there. If you would succeed, you need explicitly disable deduplication and compression on this drive, then it can be used even on Linux (however it will work slower at least 2 times than on Windows and almost 4 times slower than ext4 and will produce errors during writes, which requires Windows chkdsk to fix them, so better do not use NTFS under Linux ever).

2 Likes

But why you have MBR for a big ass 16TB disk? and why you use USB connection for such a task? Why Windows? Why NTFS? So many questions. For data recovery I recommend EaseUS.

1 Like

it’s GPT as far as I understand

and they tried to use

Well I’m confused now. Anyway it’s pointless to do the data recovery in this case. 5% loss and you’re dead.

1 Like

I have to agree. If the author doesn’t sure what geometry he was using, it’s impossible to suggest something.
Right now it looks as completely messed and corrupted. I have doubts that it can be recovered above 96%

2 Likes

Wasn’t changing anything in geomethry.
The 16TB HDD is NTFS with GPT mode, not MBR. (because MBR is only for up to 2TB)
First program i attempted to fix it, was TestDisks.
TestDisk couldn’t fix anything!
But TestDisk showed MBR is Ok.
How ever TestDisk said filesystem is corrupted or damaged, couldn’t read files or find MFT,
and couldn’t do anything with this.

The HDD was orginally working under Windows 7 Pro.
Where it was passed to virtual machine, where node was under Win10.
Passed only with that 14,55TB partition, where the node was.

The disconnection was made as fallowed:
Stopping the node, waiting some minute, turning the virtual Win10 off.
After this under host Win7, the disk reappeared again with files shown normally.
Then safe removal icon was clicked, and disk unmounted, because it was connected via docking station (clone device) via USB. (was working like that for 3 months, it was temporary)
And WRONG move here: i took the HDD from the docking station, while it was still power ON, and connected physically to PC. Didn’t want to power it off, because was afraid of changing disks orders under system, and under virtual machine could cause problems for rest of HDDs.

The cause is known: improper unplug from the PC (just like with power outage case)
It’s not a more mess that it was.
The screens from different programs just shows it from perspective.
None of them was able to do anything, nor even have available option to press to fix!

Only Gparted, had, BUT even ran with system-secure.org version,
shows it can’t do anything, and showes it has support for NTFS:

Geparted just shows it can’t open $MFT.

In addition i read about that ntfs 3g and tryed to instal itt or mount but no use:


leading to same conclusion, he can’t read MFT.

Great so i found a video, about fixing a RAW

indeed Windows doesn’t see a HDD under any filesystem, for windows its RAW,
just blank, just like on this video.
The program it suggests is free minitools Partition Wizard 9.
Well todays it’s ver. 12.8, i installed it, in trial of a full version,
it is a breakthrough, because it’s the only program that’s able to show what files are IN,
(just like suspected, there are some private backups)
BUT, next step “Check File System”, as advised,
it can’t pass above a certain step, and just stops,
just like an original chkdsk before PC boot.

So what, to put the HDD back on docking station?
as it is still turned ON, and try there, maybe original windows remembers something about MFT?

please never do that. They are not backward compatible.
Seems your disk is affected by influence of Win10, so you cannot (and shouldn’t) use it under w7 anymore.

Perhaps it can recognize the drive, but after usage of different tools…, I’m cleanly. not sure…

Thx for insights. How ever i was using successfully disks from Win7 host, passed to a Win10 in virtual machines for years, it always works back in win7, file system always working on Win7. Probbaly because its created under Win7 in first place.

Yes. If it has been formatted under Win10, it will use new features by default. It also uses them for the existing drives, formatted earlier (I.e. in WinXP, Win7, GParted, …), when you use this drive to store data.
So maybe Win7 has these features too, I cannot check anymore, I do not have any Win7.

That should not destroy your GPT.

Only me see there is a contradiction?

Sigh…

In that specific case you must off both systems, the VM and the host.

Unfortunately in this specific case it is much more disruptive, since the partition was shared to the VM.
GParted can work with NTFS via ntfs-3g package, which is not fully supports NTFS… So, it cannot fix anything. Especially after win10.

I have a strong feeling - it’s dead.

2 Likes

Not only You, i see too, TestDisk dissapinted me, that what it showed me, im not halucinating.
How ever im going to recover node’s data with a program that sees the files, and just repair the HDD and place the data back again, its 7th day of nodes offline, i should have 5 more days to put node online to save it! - just dont know if that will be enough time to recover, and copy it back, its 2,35TB of small files records shows.

1 Like

Have you tried to run chkdsk X: /f on this disk when connected to the source computer (to the one where it was originally connected to)?

1 Like

No. Will try maybe on source computer, but tomorrow, after i recover and secure the data before.

1 Like

I understand your willing to save the data. Maybe you should try R-Studio. It’s pretty good. But don’t expect too much. And please next time make the disk internal, if your pc doesn’t have any SATA port left then you can try PCI-SATA card, it’ll give you some SATA ports. Then use Linux where you can. I see you have technical background so it’s not hard to switch to Linux and play with KVM (since you’re familiar with VM) or Docker. It’s faster, more stable, scale up, long term…blah blah… Good luck man.

1 Like

its scanning chkdsk under win10 now, and it goes further with repairs than under Win7,
but it will take ~74h for full chkdsk as it is moving only with speed of ~7MB/s,
disk respons eis good, 1,7-2,5 ms on avg. Disk is fast, just the scan disk can’t use multiple threads to go faster i guess + 14 milions of small files over 3TB.

I saw somewhere here that SNOs can submit problems with a node,
for some additional halting (like audits?) for a node
if there are some procedures, i would like to ask for some halting for this node: 12mMKmuNMzBx5uhz9MoQU7Eddr3tN5wkyqoidoYkzKVMCAiTsrv

because the recovery is undergoing!
If There could be no more audits if possible, untill it will come online, that would be great!

1 Like

Your node is like an alcohlic driving 180km/h on the motorway, with only 2 hours of sleep, run down tires and heavy snow. It can work for some time, but it will not work for long. Now you are asking the police officer to give you a second chance and let you continue your drive because you will sober up eventually.

Give up, learn from it, rebuild it proberly!

3 Likes

What are You talking about, stoooop.
The files are perfectly healthy, and The node was perfectly healthy for 3,5 years.
It’s just the access to partition i broke.
That’s not even funny.
Thank You, for Your advice’s.
if i make it in time, the node will be healthy for more years to come.

1 Like

I ran out of popcorn watching this thread, you guys are hilarious!

This shall be turned into a masterclass “Heavy artillery: practical advice on aiming at ones own foot”

Sorry, I don’t have anything constructive to contribute: the whole things needs to be bulldozered down to the bedrock, as there is no point in fixing this madness, and instead a new architecture needs to be implemented – without Windows, without raid controllers that are not in IT mode, and without the unnecessary VMs…

4 Likes