1 (edited by jgkurz 2016-12-14 18:08:02)

Topic: SOLVED- GParted - fsck Progress

Hi folks, I have a situation I hope you can help with. I did a stupid thing and incorrectly replaced a bad drive in a RAID5 array (LSI 9260-8i) . I didn't remove and replace it correctly and probably corrupted the stripe. I went back and replaced the drive correctly but the files were not accessible afterwards. I decided to boot systemrescuecd and run an fsck check from GParted GUI before I consider the data lost. I'm not much of a Linux admin which is why I used the GUI. The file system is ext4 and just under 12TB so I expected the fsck to take a long time. I'm at 36hrs and it's still chugging along. The task manager shows the PID using ~50% CPU with RSS around 200MIB. The command the GUI ran was "e2fsck -f -y -v -C 0 /dev/sdb1"

There was a progress bar but when I opened another window it stopped refreshing and is blank. I'm guessing it's because CPU is busy and not refreshing the screen.

1) I can let the fsck run a week or more if needed. At what point do I kill the process?

2) Would killing the process and restarting cause further corruption?

3) Is there a CLI command that could give insight into the current operations without killing it?



Thank you!

2

Re: SOLVED- GParted - fsck Progress

Unfortunately it appears that you are complicating the issue by performing various steps before getting an understanding of the situation.  Each step further reduces the chance of recovering data.

I recommend that you read up on how to correctly replace a RAID5 drive, instead of running fsck on whatever remains in the existing hardware enclosure.

You may have to revert to your backups, depending upon your comfort on researching what the current state of things are.

With respect to your three questions, I can't really say if letting fsck run or killing it now is the better option.  As for CLI commands, you can look into ps, and iostat.

3

Re: SOLVED- GParted - fsck Progress

gedakc wrote:

Unfortunately it appears that you are complicating the issue by performing various steps before getting an understanding of the situation.  Each step further reduces the chance of recovering data.

I recommend that you read up on how to correctly replace a RAID5 drive, instead of running fsck on whatever remains in the existing hardware enclosure.

You may have to revert to your backups, depending upon your comfort on researching what the current state of things are.

With respect to your three questions, I can't really say if letting fsck run or killing it now is the better option.  As for CLI commands, you can look into ps, and iostat.


Thanks for the response. I think dug my way out of this mess. The killed the fsck started from the GParted GUI and then restarted the fsck from the command line. It finished in a few minutes and corrected a handful of errors. I re-ran it again and it finished with no issues. Next I started my NAS and mounted the file system. All my data was there! So far I can find no corruption. Next time a drive fails I will have a better game plan.

4

Re: SOLVED- GParted - fsck Progress

Thanks jgkurz for reporting back with how you resolved the issue.

To help others searching for answers to similar questions you can edit your initial post and prefix SOLVED in front of the title.