Oct 27, 2017 It is suggested that you go to Storage Manager to run 'Check File System'. But I can't find this 'Check File System' anywhere - can someone show me where this is? TS-879Pro (firmware 4.2.0, 8 hard drives in a RAID6, accessed from Apple machines). There's a risk of filesystem corruption in case of system crash. And when checking the datavolume, the total size is clearly wrong. When running e2fsck (or e2fsck64) it freezes somwhere in Pass 1: # /bin/e2fsck -C 0 -v /dev/mapper/cachedev1 e2fsck 1.42.13 (17-May-2015) DataVol1 contains a file system with errors, check forced.
We got this QNAP TS-859U+ with firmware version 3.8.1 Build 20121205 at our datacenter. It has Intel(R) Atom(TM) CPU D525 @ 1.80GHz processor and 1GB RAM, 8 of 3TB (Seagate ST33000651AS CC44) drives and they form a 7 drive RAID5 array. The other disk is a global spare.
My intention is to recover as much data as possible.
After a power failure, there was this log message:
[RAID5 Disk Volume: Drive 1 2 8 4 5 6 7] The file system is not clean. It is suggested that you run 'check disk'.
That RAID5 logical volume was still mounted and we had the chance to start a filesystem check from the QNAP Web GUI. But we decided to do this after work-hours not to cause any inconvenience to the users. But we never had the chance again because the device rebooted itself and the RAID5 logical volume became 'Unmounted', so it wasn't possible to start a filesystem check from the GUI anymore since the 'CHECK NOW' button became inactive.
I started 'Bad Blocks Scan' for all drives and they all completed successfully. They all say 'GOOD' for SMART information.
Then I tried to mount that volume manually via SSH and this is the output:
This mounting attempt's reflection on dmesg:
Here is a longer dmesg output from device startup:
I checked and the RAID is active for md0:
Superblock is persistent as well:
I tried various e2fsck_64 (even e2fsck_64_qnap) command combinations like:
..of course after the 'adding extra swap' ceremony, because it quickly throws a 'memory allocation error' otherwise:
The scan hangs like this:
If I use e2fsck_64 -p, it also adds a CLEARED. message at the end of the line. But it doesn't go any further. Meanwhile, e2fsck_64 process' CPU usage drops to ~0,9% but it still uses around %46 memory. I doesn't look like it's making any effort. System RAM is almost full but it seems like it no more fills any swap space.
I tried adding a USB stick as a bigger swap as the user RottUlf described here: http://forum.qnap.com/viewtopic.php?p=216117 but it did't change a thing.
I also created config file at /etc/e2fsck.conf like this:
..and used a USB stick for that purpose:
..as mentioned here: http://forum.qnap.com/viewtopic.php?f=142&t=102879&p=460976&hilit=e2fsck.conf#p460976
It didn't help either.
Some documents recommend trying to run e2fsck_64 with a backup superblock but I couldn't find any:
Lastly, I tried to recreate raid with mdadm -CfR --assume-clean because I've read that it helped some people out there who experience similar issues, to get their volume mounted and see their data so they can backup:
..but it didn't help, still can't mount, same errors.
We also have a beefier QNAP, model TS-EC879U-RP with firmware 3.8.4 Build 20130816. It has around 3.76 GB usable RAM and Intel(R) Xeon(R) CPU E31225 @ 3.10GHz processor. But it's completely full with another set of important data.
So, what I have in mind is to shut the both QNAPs down and take all 8 disks out marking the slot order, keep working QNAP's all 8 disks in a safe place, and put TS-859U+'s disks on TS-EC879U-RP with the correct order and run e2fsck_64 on that powerful QNAP. But I don't know if the other QNAP will correctly detect the problematic RAID at 'Unmounted' state...
..or the data on the powerful QNAP will be retained after it ever manages to finish e2fsck_64'ing the 'guest disks' and I put all the disks in their original slot and power on.
Any help will be greatly appreciated,
Thanks in advance..
ubunmutuubunmutu
2 Answers
The order of the disks won't matter, the configuration for the RAID is stored on the controller, which is in your older system and moving the disks to another controller will just present 8 new disks for it use. It won't know about any existing data.
Was the file system encrypted or just a standard RAID 5? Use RAID 6 next time :)
0x0000001E0x0000001E
I had managed to recover nearly all the data with the help of TestDisk, after connecting all 7 drives to a PC. TestDisk had managed to detect the broken file system on the RAID5 volume and export most data intact.
ubunmutuubunmutu