Errors with ext4 HDD, umounting

[nezhac]

Hey all,

I've been having some issues lately with the HDD connected to my pi. I've got the tenda wifi dongle from modmypi (https://www.modmypi.com/tenda-wireless- ... i%20dongle) on one of the pi's ports and the other connected to a 4 port hub (https://www.modmypi.com/New-Link-4-Port ... ns-Adaptor). I Recently switched to a separate power plug for the pi and the hub, so they're independently supplied, in an effort to isolate the problem.
Here is an output of the error messages I get
$this->bbcode_second_pass_code('', '[38917.844557] EXT4-fs error (device sda1): ext4_find_entry:1309: inode #20733953: comm transmission-da: reading directory lblock 0
[38917.861104] EXT4-fs error (device sda1): ext4_find_entry:1309: inode #20733953: comm transmission-da: reading directory lblock 0
[38917.879094] EXT4-fs error (device sda1): ext4_find_entry:1309: inode #20733953: comm transmission-da: reading directory lblock 0
[38917.895820] EXT4-fs error (device sda1): ext4_find_entry:1309: inode #20733953: comm transmission-da: reading directory lblock 0
[39110.912093] EXT4-fs warning (device sda1): __ext4_read_dirblock:908: error reading directory block (ino 2, block 0)
[39119.360278] Buffer I/O error on device sda1, logical block 121667584
[39119.371882] lost page write due to I/O error on sda1
[39119.372192] JBD2: Error -5 detected when updating journal superblock for sda1-8.
[39165.467193] EXT4-fs (sdb1): recovery complete
[39165.468442] EXT4-fs (sdb1): mounted filesystem with ordered data mode. Opts: (null)
')

I've noticed it seems to change the identifier from sda to sdb when I try to remount the drive after. If I try to look into the HDD after this error I get
$this->bbcode_second_pass_code('', '[robin@alarmpi ~]$ ls /media/HD-PCTU3/
ls: reading directory /media/HD-PCTU3/: Input/output error
')

Here is some more info, kernel I'm using, my fstab and the results of e2fsck (which seem to come back clean)
$this->bbcode_second_pass_code('', '[robin@alarmpi ~]$ sudo e2fsck -vfp /dev/sdb1
HD-PCTU3: recovering journal

32382 inodes used (0.05%, out of 61054976)
4692 non-contiguous files (14.5%)
20 non-contiguous directories (0.1%)
# of inodes with ind/dind/tind blocks: 0/0/0
Extent depth histogram: 29762/2584/28
197201373 blocks used (80.76%, out of 244190632)
0 bad blocks
9 large files

29782 regular files
2591 directories
0 character device files
0 block device files
0 fifos
0 links
0 symbolic links (0 fast symbolic links)
0 sockets
------------
32373 files
[robin@alarmpi Music Videos]$ uname -a
Linux alarmpi 3.10.28-1-ARCH #1 PREEMPT Tue Jan 28 15:46:05 MST 2014 armv6l GNU/Linux
[robin@alarmpi ~]$ cat /etc/fstab
#
# /etc/fstab: static file system information
#
# <file system> <dir> <type> <options> <dump> <pass>
/dev/mmcblk0p1 /boot vfat defaults 0 0
UUID=c02521c7-0798-4e11-acb7-548b8ac0a4a1 /media/HD-PCTU3 ext4 user,auto,noatime,nofail 0 2

')

I run transmission-daemon on my pi, and the error tends to appear after I leave my pi just running transmission for a few hours. I come back and try to ls the hdd but get the above error. It has also happened a few times randomly while I'm watching something with xbmc, the video abruptly stops and the hdd disappears.

Thanks for any help, if there's more info I can supply to help just ask.

[moonman]

is your hdd independently powered? i've had problems with hdds powered through the usb hub even.

[nezhac]

No, I'm using a 2.5" external so it's powered through the hub. Do you think the hub is having issues powering it?
That was my thinking behind getting an extra seperate power adapter for the pi, so the hub would only need to power the hdd. What I'm also worried about is if my hdd is failing and going to crash soon.

Thanks

[sdjf]

I wonder if the speed of your SD card might also be an issue.

I don't have HDD experience, but have multiple memory devices plugged in to my pi through a usb hub, and find that when the Pi goes into a snit, unable to read what it should see, I umount everything and sync and reboot and then all is well. Not sure how you would prevent that with something that needs to keep running all the time.

[nezhac]

sdjf, do you mean the speed of the SD used in the pi itself? I'm currently using a sandisk 32gb class 10 so I don't think that's the problem.
I was recently messing around with the cables and stuff and noticed the HDD got unmounted when I was twisting and moving various usb cables, so I think the issue was maybe a faulty connection. Rearranged the pi and HDD so there's not so much pressure twisting the various connections and so far I haven't had any unmounts or I/O errors, as it's a sporadic problem I'll wait a little longer to see if it comes back.

Thanks