Partitions lost without warning

Gparted in Ubuntu. And i knew creating a partition on a troubled disk was harmful, but i did it after failing at testdisk, gparted recovery and RStudio. It was to see if the drive itself had problems.

NTFS within Linux i.e. gParted is reverse-engineered. When you want to have an NTFS formatted disk/partition you should always create it within Windows, as it is obviously using the actual NTFS.

Well, I have reasons for both, but I certainly do not know enough to have a strong opinion!

Reasons to think it was hardware?
The previous drive failed with bad sectors in the beginning. Later, I re-partitioned that failing drive thus:

  • 30 GB unallocated (to coincide with the known bad sectors)
  • 100 GB NTFS, for just a small amount of stored documents (old ESL textbooks, etc)
  • one Manjaro installation, 60 GB
  • One Mint installation, 80 GB
  • the rest was Linux Swap.
    Since nothing here was essential, I was fine with running a failing disk. It ran with no errors, only warnings at startup and booting into Linux beautifully for several months.

In early June I bought a “new” harddrive and did the folowing:
sda1, 283 GB, and filled it with the music, video and old textbook files, total 160 GB
sda2, re-installed Manjaro, same size as before
sda3, installed Deepin, 100 GB
Linux swap to end, 35 GB
Two weeks later, I removed Manjaro and Deepin, merged them and installed Mint.
One week after that, the partition was gone.

Reasons to think it was software?

  • This Ubuntu installation has added Mate Desktop to a Gnome installation
  • Those three missing Mate desktop libraries
  • There have been no other signs of hardware issues, other than some annoying ways that USB devices get dismounted (very annoying when the camera is not found in the middle of a tutoring session).

Tomorrow, I am bringing it into a shop almost entirely staffed by serious Linux geeks. My patience (and available time) is gone.

i hope the shop is able to give you some clarity about the best way to move forward :slight_smile:

2 Likes

Took the words out of my mouth @01101111. I too hope you get a solution @cliffsloane and you can tell us what it is when you do.

2 Likes

Well, friends, I just came back from Re-PC. I spoke with 3, count 'em THREE, serious Linux geeks. The lead geek looked like he came straight out of Central Casting – big bear of a guy with a long beard and even longer ponytail. He was very kind, spent a lot of time with me, but they all said largely the same thing:

“I dunno. That’s never happened to me.”

Best guesses, in order of likelihood, were:

  1. Windows is up to no good. But that is nearly impossible because even Windows 7 warns you before it changes most things, and would NEVER just delete a partition table.
  2. A bad drive. But badblocks, fsck and SMART would all reveal that, and they did not.
  3. Some issue with Ubuntu. But none of them ever had missing library files, even when switching around DEs, so that seems like grasping for a single straw.

So going through logs wouldn’t help because even they, the high-powered geeks, would not know what to look for.

Here, then, are the feeble and seriously unsatisfactory steps I will take:

  1. Install Windows 7 on either a VBox or, following @abhishek’s tutorial, on a USB 3 pendrive;
  2. Format the Windows partition (/dev/sdb2) and use it for storage
  3. Install Mint on sda and run the disk health utilities every week.

I feel like I do when I go to the doctor, and am told there is nothing wrong with me, despite the aches and pains.

2 Likes

I am surprised you present such steps. You actually don’t have to fiddle around a lot if you just do a solid backup every week. Besides that, no fiddling and setting up can replace a good old solid backup.
So

should actually be: run the system backup every week!
(Especially, since, as mentioned in previous posts, these “health” utitlities show the status quo. So… if your disk is broken and it lights up red in those utilities, it would be like… The train has already left the station as we say here where I am.)


This statement is also not true in all cases.

In the “Train Has Left the Station” department, the best advice I got, in addition to running weekly backups, is to create an image of the troubled drive and work on the copy, not the real one.

Panic and anxiety have a way of clouding one’s judgement, yes?

1 Like

"Do not panic! " is probably one of the phrases that makes people panic the most. “Stay calm!” is hard to follow if you’re looking at a desaster with your own eyes. It takes professional training and an extreme amount of practice to get along (relatively) fine in such situations.

In itself the sentence is entirely correct, but in the context it’s not really correct, as it implies you have to fiddle around with a troubled drive, in the first place. Don’t fiddle, just use your solid backup, that’s enough. This technique of taking an image is standard practice among digital investigators and data recovery specialists, but it’s more or less still just the entry point of where you have to know the other tools, as well… Imagine you have an issue with the drive, take an image and then what? Well, you could fiddle around with it but if you don’t know what exactly to fiddle you might still fail and it wouldn’t help much.
I see that I state simple things in a way too verbose manner now. All I want to say: keep SOLID (i. e. tested, proven to be working, tested, you being absolutely sure they work, they are based on a very redundant backup mechanism, etc. and of course tested, tested, tested) backups and your job is done. Just back up. That’s it. That is pretty much all you gotta do. Press the button, to back up. Press the button, to restore.

Don’t waste your time, just push the button.

1 Like