Corrupted filesystem soon after a fresh install


#1

Hi,

Thanks to anyone that can help me out of this daunting situation !

My PC has 2 hard drives, all filesystem but swap are formatted in ext4 :
-One SSD where is located a Ubuntu 14.04 partition (sda3) that I have been using since 2014 [sda2 was a swap partition that I deleted when I learnt that shouldn’t go on SSD and (sda1) must be some kind of boot for ubuntu, I forgot]
-One HDD with 3 partition : (sdb1) which I use for storing all my personal files and which I always mount on /data (in order to avoid mixing /home configuration files and my personal files) + (sdb3) where I have installed last week a fresh Manjaro 16.10 Xfce + (sdb2) swap

The process of installing and configuring my favorite packages took me about 3 days during which everything was going perfectly fine. Appart from a driver for my Samsung clx-3185 printer which was downloaded from Samsung website at day 1, all the installs and remove packages was performed with pacman and pamac. Powered off and back on a number of times without problems.

Yesterday evening , noticed that Firefox was not opening and decided to power off and back on. That’s when problems started to appear in the below order :

  •    Screen blocked at the splash screen “Manjaro ….”  --> used the Pc power off button to leave and reboot without quiet splash
    
  •    At reboot : system advising very quickly on boot that filesystem was corrupted and needed a fsck without the –p or –a option "unexpected inconsistency"
    
  •    Rebooted on the Ubuntu 14.04 to perform the fsck which found numerous disk problems (asked to extend lost+found at some point)
    
  •    Gave on other try at Manjaro which went a bit further this time until mouse and blinking cursor seemed to enter in an endless loop on the screen, tried ctrl+alt+F1 and read something like
    

failed command : READ DMA EXT

blk_update_request : I/O error, dev/sdb, sector 1844813048

  •    Went back an other time to Ubuntu and performed an other fsck which fixed a different set of problems (some looked similar but others was definitely new)
    
  •    Just to be sure ran fsck an other time but with the –p option and it said that the filesystem was clean (lost+found seems heavily populated but folders on the root of the tree are there)
    

After reading the different posts, I think I should now use mhwd-chroot as explained in https://forum.manjaro.org/t/how-to-save-your-manjaro-installation-when-it-breaks/75 to try and repair the whole thing but two questions come to my mind :

1 - What to do once I will have successfulyy chrooted the system ? Is there a command to check what is broken or to re-install every single package of my system ?

2 - No idea if the problem is coming from hard or soft, smartctl says my disk power on time is abt 30 000 hours but most of it being idle as I never stop the PC but wondering why did it happen just after installing Manjaro, is this something frequent here ?

thank you very much, here below some additional infos on my system :

sdb = Western Digital Scorpio Blue Serial ATA (AF)
cpu = Intel® Atom™ CPU D525 @ 1.80GHz - cores=2 enabledcores=2 threads=4


#2

Sounds mysterious. I don’t think chroot can help. Did you run a SMART test? Edit: I see you did. No idea.


#3

What does smart say about Reallocated_Sector_Ct and seek_Error_Rate and Raw_Read_Error_Rate ?


#4

download this from their site build in ubuntu its easy to do just unpack in a folder same name follow install instructions when built run from terminal follow the read me don’t bother with the smart tests as you have done them go straight to hdrecover /dev/hdx if ide, /dev/sdx if sata change the x to drive letter let it run it will try to recover bad corrupted sectors if not recoverable will delete that sector you should end up with a usable drive if its not to far gone till you can replace it, saying that I recovered a 2tb disc damaged from not powering off correctly 2 years ago reporting 200 bad sectors smart test the customer is still using it daily smart tests say the disc is in good health no bad sectors?


#5

@ mandog:
Can you please rephrase this using correct interpunction and correct capitalisation to make it understandable? What is this “this” you are proposing to download anyway?
J.


#6

Thank you all !

@Jsamyth : smartctl output

Vendor Specific SMART Attributes with Thresholds:
ID# ATTRIBUTE_NAME FLAG VALUE WORST THRESH TYPE UPDATED WHEN_FAILED RAW_VALUE
1 Raw_Read_Error_Rate 0x002f 199 199 051 Pre-fail Always - 500
3 Spin_Up_Time 0x0027 228 211 021 Pre-fail Always - 1591
4 Start_Stop_Count 0x0032 100 100 000 Old_age Always - 212
5 Reallocated_Sector_Ct 0x0033 200 200 140 Pre-fail Always - 0
7 Seek_Error_Rate 0x002e 200 200 000 Old_age Always - 0
9 Power_On_Hours 0x0032 056 056 000 Old_age Always - 32444
10 Spin_Retry_Count 0x0032 100 100 000 Old_age Always - 0
11 Calibration_Retry_Count 0x0032 100 100 000 Old_age Always - 0
12 Power_Cycle_Count 0x0032 100 100 000 Old_age Always - 211
192 Power-Off_Retract_Count 0x0032 200 200 000 Old_age Always - 81
193 Load_Cycle_Count 0x0032 200 200 000 Old_age Always - 176
194 Temperature_Celsius 0x0022 107 089 000 Old_age Always - 43
196 Reallocated_Event_Count 0x0032 200 200 000 Old_age Always - 0
197 Current_Pending_Sector 0x0032 200 200 000 Old_age Always - 36
198 Offline_Uncorrectable 0x0030 100 253 000 Old_age Offline - 0
199 UDMA_CRC_Error_Count 0x0032 200 200 000 Old_age Always - 0
200 Multi_Zone_Error_Rate 0x0008 100 253 000 Old_age Offline - 0

I thought that since there are a lot of file in lost+found, I would need -even if I successfully repair sdb3- to restore these system files to their original positions. No idea what are these files but thought I could find out by booting into the system with chroot and in the worst case use pacman to list ALL packages installed and reinstall each and everyone.

@Jeannie and @mandog : yes I believe mandog slipped the name of the utilitary he is suggesting, would be great to know :wink:


#7

And you’re not using spotify?


#8

Hi, talking about a music streaming website ?? No I am not using it.


#9

Yes - just speculating :slight_smile: - since spotify screwed a lot of storage space around the world


#10

I would like to be sure that Manjaro is not the guilty party actually since I was already in love with her… and to know in such case what have I done wrong…


#11

Sorry hdrecover is the program not really hard to work that out unless you choose to be politically correct of course. :imp: http://hdrecover.sourceforge.net/ By the way jeannie you know very well that i’m dyslexic so no need for personal attacks on me.


#12

I am quite sure its now manjaro - this will happen sooner or later - I think you’re off to get a new drive.


#13

yes you have to know what happened but I doubt if its Manjaro that caused the problems. bad shutting down can cause these problems, bad power supply internal or external win10 is also a culprit as it does not always unmount the drives, bad connections is another culprit as is bad ram.


#14

Still your post is unreadble.
J.


#15

Thanks, once again I dont want to blame Manjaro but no windows has ever been installed on this PC. Bought it with ubuntu 10 which ran continuously during about 3 years then ubunutu 14 which ran continuously during more than 2 years and just 3 days after switching to Manjaro, the very partition of Manjaro failed. Strange coincidence ! I am sure that last switch off was a proper one with the desktop icon. Trying to figure out what I have done wrong , my only idea is recalling having used ctrl+c on a long “du -kh” command which at second thought was not the command that I needed ( I use this type of commands once every 2 years so my personal memory is also questionable :frowning:)


#16

Btw is ext4 the recommended file system for Manjaro or is btrfs deemed stronger for my next install ? any idea how to quickly “clone” the system by reinstalling every single package (without removing the configuration files that are still there) ?


#17

Btrfs is not yet ready as default in my opinion as not all the features are usable ext4 is a very capable system use clonzila for backups it looks daunting but at a second look very easy to use


#18

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.