A content of the entire system partition has been deleted

But, it does happen.

1 Like

I agree with what tbg says, and that is why when installing any OS with any installer (not just calamares) we will need to verify the target partition with one of lsblk -f, blkid, fdisk -l … and when doing repair (say grub-repair) or chroot, again this verification must be done on each reboot. Often times, I have to repeat the adage that it is the bios that determines what is sda, sdb… not the user; and that changes with each boot.

However, while I agree with tbg’s point, it cannot be conclusively said that this is the cause of the OP’s problem. At most, it can only be a conjecture. But we can accept that it can be a possible cause.

This isn’t helping much, but there’s not much we can go on anyway.

But cheers, as always. :slightly_smiling_face:


When we compare our cases more and more factors seems to cover, though I am not able to say which one caused those damages.

One Linux geek suggested me that maybe this issue has something to do with a very strange Linux kernel “feature” of udev, in which, in some cases, “if the machine has two hard drives, /dev/sda may randomly become /dev/sdb”. And we both had two drives…

On the other hand, we both use ext4. I realize that ext4 is well tested, but sometimes… “sh.t happens” (but why at my place :rage: :smile: ) .

You didn’t lose your partition’s data because of ext4. You lost your partition’s data because you did something to the partition.

What were you doing on your system two days ago?

1 Like


I am not saying that for sure problems were generated by ext4 or something else, as I do not know it. I just see some similarities between my case and that of @AgentS.

I can not exclude that in my case, data lose on sda5 (2 days ago) was some consequence of kernel 4.9 issue (7 days ago) or of a serious GRUB issue (5 days ago).
And 2 days ago, as I wrote in the first post, I have not done anything related to the system or sda5. I only checked Manjaro Installer (though again I want to emphasize that I do not know what was the main reason for my data loss on sda5 2 days ago).

Well, there isn’t much more we can do. I also noticed device’s names changing at boot. That’s why UUID’s are better then /dev/sd designations. If the partition was formatted and not overwritten it can be recovered using a tool (in ex.: testdisk). Otherwise just forget it, maintain backups and, if you want to clarify what happened, try to replicate the problem. Formatting drives can be tricky. It is always a good idea to double check what you’re doing, like @gohlip said.

In that 2nd link (post 18), you said you chrooted into Manjaro.
How did you chroot? mhwd-chroot/manjaro-chroot or the ‘traditional chroot’?
Don’t forget you have 2 Manjaro’s and if you use these methods, you may have chrooted into the wrong manjaro (sda5). And did you mount $esp before chrooting?
Furthermore, even if you have chrooted properly, did you exit after doing the stuff. Did you continue other operations thinking you are in livecd OS when in fact the operations apply to that Manjaro installed partition? [That’s why I discourage people making chroot, if possible]

And yes, these are possible causes, and you may have done everything properly.
But the fact is that most times (especially by ‘experts’) we make mistakes without realizing it because we thought we know we did such simple things correctly.

Again, not saying you’re wrong or made a mistake (we can’t tell, even you can’t tell).
We just don’t know. I think we should just put this episode down as an open case and move on.

ps: I myself have made many mistakes (chroot, dd, more…) and in a few times don’t even realize it.

Luckily for me, I had nothing important on that cleared partition, so I do not need to recover anything from this partition.
What is important to me, I would like to discover what was a most probable cause of this issue, just to eliminate potential risk factor and use the system with confidence.

Answering you: I was using sudo mhwd-chroot-shell command and after I got a prompt to choose the system partition, I chose number 0 (from: 0 = sda2; 1 =sda5; 2=sda7). And I am sure I used exit command after doing the stuff. After all I rebooted.
And after I restored GRUB from sda2, I was using the system (on sda2) for ca 3 days, without any issue. So IMHO, it is less probably, that this GRUB issue was a direct cause of data loss few days later.

At the end of the you are not going to admit if you made a mistake or not, this is 2018 it is always something else or some one else to blame.
What I will reinforce is in 30 years never has a installer on win/Linux made any decisions that I did not authorise before hand it is not possible for that to happen, we make the decisions. not the installer not chroot not the keyboard, we mount the partitions, we yes we make the mistakes, we conveniently forget what we did and blame it on the tools.
Every user has made mistakes I did just what you did a few years ago and deleted 500gb of data, how did it happen well its called human error I was trying to rush the phone rang next thing all data was lost, now I could blame the phone, I could blame the Wife, I could blame the partitioning tool but no it was human error simple as that no user is infallible, not you not me, not anybody, Just accept what has happened learn from it
not blame everything else for what is most likely a simple human error.
The best way to fix it is like this


You hit the nail on the head there @mandog. Everyone’s done it, but nobody likes to admit it. Hell, I flushed 4 TB of data a few months back with a bad dd command. I was in a rush and i was flustered, and sh*t happens sometimes.

My heart dropped when I realized what Id done. Then I shrugged it off because I had another mirrored 4 TB backup at home, and another off site. Backups turn those moments from a catastrophe into a mere inconvenience. Backup, backup, backup, it can’t be repeated enough.


The problem is that in this case I do not remember doing something incorrect :left_speech_bubble:
I realize that many problems are simple human error and please believe me, that if I would make some mistake, I would admit it, as I did in the second sentence of this thread, writing: “I mistakenly installed the second GRUB”.
And here we have 3 similar cases (my, @AgentS, @emerilin), where everyone:

  • was using Manjaro Installer (in different versions),
  • had 2 internal disks,
  • used ext4 (at least me and @AgentS ).

Perhaps this is a pure coincidence, but perhaps a composition of some of these three factors builds a (rarely occuring) explosive mixture :bomb::boom:

Highly, highly unlikely.

You can test this by running a VM with the same setup, then trying to replicate the same steps as you performed several days ago.

Of course, given the amount of formatting, re-partitioning, and OS installation you’ve done it’s going to be very difficult to replicate.

If you’re concerned there’s a lower-level issue you might consider wiping the machine and starting from a clean slate. That way there’s no left-over configuration which may get in the way or produce spurious, unreplicatable issues.

Plus you have less stuff to blame.

1 Like

I totally agree about wiping the drive. Totally wipe the partition tables and their backup entries with a dd command. You can reinstall what you want back via imagining. That way if you ever have to recover data again you don’t have a dozen old partition entries to confuse your recovery software. You should also get in the habit of backing up your partition tables. That is the surest way to get your data back with the proper file names.

You really need to simplify things. What your doing now is causing you nothing but grief. Go to VM’s if you insist on running a bunch of different OS’s at once.

Hey, careful there. I (and wongs) have lots and tons of OS’s , nothing from VM’s.
Don’t blame it on the system when it’s the user outside the system. As you said, " nobody likes to admit it. " Ha ha.

ps: I wouldn’t have written this if you are ‘other people’. They would have taken offence. Not you.
Is that a compliment? Nah! Ho Ho Ho!. couldn’t resist. Ngah Ngah Ngah!

I personally disagree , but take no offense. Running multiple OS’s on metal is looking for trouble unless you are extremely knowledgeable. Even then Microsoft will likely throw you a curve ball at a future upgrade and bork your system. The people that really know their stuff might get away with many OS’s with out major issues, but the average user will probably have big problems.

@gohlip may be the sensei when it cones to fixing borked boot issues, but look at how many help requests these multi boot systems generate on the forum. One OS per hard drive in my system. I use a hot swap rack and just run things clean and easy. If you play with fire you’ll likely get burned.


I have 5 hdds all running at least 12 systems on my desktop, i do a lot of partitioning the drive orders can and do change at boot time nothing to do with Linux, or partition managers. its a feature of sata drives they are all equal and no boot order that is why you have to use UUID to identify partitions not driove boot order.

I can honestly tell you even I make mistakes unknowingly after 30 years and with a memory that is old i do not remember what I did wrong, but i’m the operator everything is my responsibility not the machine it only does what I tell it to do it has no brain partitions can only be formatted if the operator mounts them 1st or changes their mind and forget to unmounted them, always if you change your mind reboot start again from scratch.

You are very correct on that point Users tend to forget how complex it can be 2 days ago i did a read test on one drive then i found i had done a random read/write test, Luckily it was the win 10 partition so no big deal. Luckily after a couple of hrs trying to reboot win10 it miraculously reached the desktop don’t ask me how, after a disc check and registry optimise win10 was saved, again i was distracted just for a milli second.
But this is 2018 should I blame the Gnome disk Utility I mean it is Gnome after all. Or was it the Sranglers music I was listening to, or divine intervention, No it was me I clicked proceed


A true (wise) word, and it should be nailed on top of every help forum.

None taken. {I seldom take offence, nothing special here :grin:}
But as they say,

It is not difficult if you know how.

or as a Chinese (Ivana Trump) saying goes… :rofl:

“Those who say it can not be done, should not interrupt those doing it.”

Or as a real Chinese (meaning ‘I’ :laughing: ) say…

He who has apples should give to those who have none.
Once given, is none of his business.

Ho Ho Ho.


There is one other possibility I hadn’t considered until now. I went through a situation 4 years ago where my GF’s windows computer had external drives in a dual bay hot swap tray that kept losing their partitions. This happened on three different occasions. This is when I started to get serious about backing up partition tables. I had done it before this but not that often.

At first I just attributed it to the fact it was Windows and NTFS and therefor a POS. The second time it happened about 6 months later I replaced the drive that had lost its partition table twice. The third time I finally clued in, it wasn’t anything to do with the drives or windows. It was the usb enclosure itself causing the problem. I assume it had a faulty controller that was causing the loss of the partitions. As long as you only had 1 drive in the rear slot of the caddy it was fine, but if you ran two drives sooner or later the first drive would lose its partition table and be unreadable.

I do not think the chances are high that this is occurring with you, but maybe you should download the phoronix test suite and run some tests on your hardware.

If it actually is a hardware issue I will be the first to offer you an apology.

1 Like

Forum kindly sponsored by Bytemark