0

After problems with mount I run e2fsck -y -b BLOCK /dev/sda and it was running for 3 days on 6TB raid 0, when it completed this was the output:

/dev/sda: ***** FILE SYSTEM WAS MODIFIED *****

/dev/sda: ********** WARNING: Filesystem still has errors **********

/dev/sda: 1529942/122060800 files (7.5% non-contiguous), 9080905/488243200 blocks

So I mounted this virtual drive and it shows now: Total 1.8TiB, Free 1.8TiB, looks like it shows the space that left on my virtual disk before filesystem ext4 crashed. None of harddisks crashed, only filesystem was corrupted. I did not have disk partitioned, only one single virtual disk.

Any way that I could restore files somehow?

Timo
  • 6,202
  • 1
  • 26
  • 28
Wiggler Jtag
  • 101
  • 3
  • http://unix.stackexchange.com/questions/82388/can-i-recover-from-a-system-disk-error-on-ubuntu-linux/. In general you need to see if you can get the filesystem on the disk "repaired" using `e2fsck`, HDAT, or a commercial too such as Spinrite. – slm Feb 06 '14 at 14:44
  • What do you mean actually by getting filesystem on? There is ext4 now, or at least it shows me in "Disk Utility" in Debian 7.1. HDAT is for windows only? – Wiggler Jtag Feb 06 '14 at 14:50
  • 2
    @WigglerJtag slm said _you need to see if you can get the filesystem on the disk repaired_ – Blacklight Shining Feb 06 '14 at 15:02
  • Well, actually there is filesystem now, but how could I get more from mountable ext4 disk without files? :) – Wiggler Jtag Feb 06 '14 at 21:31
  • 1
    This looks suspiciously like you ran `e2fsck` on one of the three volumes of your RAID array, instead of the array. What is your RAID setup? (Hardware, fakeraid or mdraid? What does `mdadm -D /dev/md0` say (or whatever is the right number)?) – Gilles 'SO- stop being evil' Feb 06 '14 at 23:04
  • No I had only /dev/sda one big harddisk, did not partition it, so it was not possible to run only on one volume. They were in virtual disk created by Powervault MD1000 as RAID 0, is that hardware or? I've seen only one 6TB disk, so could not run e2fsck on a single volume. – Wiggler Jtag Feb 07 '14 at 00:52
  • mdadm: /dev/sda does not appear to be an md device – Wiggler Jtag Feb 07 '14 at 00:59
  • @Gilles What if I've ran e2fsck on partition instead of the array, do I have any chances to get back those 2 missing partitions? – Wiggler Jtag Feb 08 '14 at 00:45
  • Maybe… but after running e2fsck on something that contained recognizable bits of a filesystem but in the wrong place, it's likely that it destroyed some important stuff. I don't think you'll recover all your data, but you may be able to recover some files. – Gilles 'SO- stop being evil' Feb 08 '14 at 00:58
  • 1
    What's the size of the disk (`blockdev --getsize64 /dev/sda`)? If it's only 2T instead of 6T then you have the wrong disk in the first place and no way to get at the missing 4T... `e2fsck -y` is almost never a good idea, not without snapshot/backup. :( But if it's RAID0 you expected complete failure anyway, right? – frostschutz Feb 08 '14 at 07:57
  • `blockdev --getsize64 /dev/sda` showed me `5999532441600`. No, I was not expecting complete failure, HDDs are working, It was like 2 weeks to move to RAID 6 and wrong system restart happened, then ran e2fsck and it showed me only 1/3... – Wiggler Jtag Feb 08 '14 at 10:42

0 Answers0