Fixdisk stuck in loop - showing "Illegal block"

Walshy

New Member
Last June, I finally got around to replacing the hard disk after regular reports of disk problems. All has been running fine until this week.
Hitting play or pause on the remote was becoming really slow to get a response, 10 or 15 seconds.
Having checked webif, I saw that I had disk errors again, so I telnetted in and kicked off a fixdisk and let it run overnight.

Checking this morning, and it's looping round and I couldn't interrupt it without rebooting from another telnet session.

Having done this, I've gone back to maintenance mode and attempted fixdisk again, but hit the same problem, with the follwing looping continuously:

Code:
Too many illegal blocks in inode 6897733.
Clear inode? yes

Inode 6897735 has compression flag set on filesystem without compression support.  Clear? yes

Inode 6897735 has illegal block(s).  Clear? yes

Illegal block #0 (2024937636) in inode 6897735.  CLEARED.
Illegal block #1 (3184809161) in inode 6897735.  CLEARED.
Illegal block #2 (4055449172) in inode 6897735.  CLEARED.
Illegal block #3 (820263210) in inode 6897735.  CLEARED.
Illegal block #4 (2962947696) in inode 6897735.  CLEARED.
Illegal block #5 (4233493169) in inode 6897735.  CLEARED.
Illegal block #6 (3914510362) in inode 6897735.  CLEARED.
Illegal block #7 (1691482175) in inode 6897735.  CLEARED.
Illegal block #8 (4257098933) in inode 6897735.  CLEARED.
Illegal block #9 (1728738978) in inode 6897735.  CLEARED.
Illegal block #10 (2320111554) in inode 6897735.  CLEARED.
Too many illegal blocks in inode 6897735.
Clear inode? yes

Inode 6897737 has compression flag set on filesystem without compression support.  Clear? yes

Inode 6897737 has illegal block(s).  Clear? yes

Illegal block #0 (2024937636) in inode 6897737.  CLEARED.
Illegal block #1 (3184809161) in inode 6897737.  CLEARED.
Illegal block #2 (4055449172) in inode 6897737.  CLEARED.
Illegal block #3 (820263210) in inode 6897737.  CLEARED.
Illegal block #4 (2962947696) in inode 6897737.  CLEARED.
Illegal block #5 (4233493169) in inode 6897737.  CLEARED.
Illegal block #6 (3914510362) in inode 6897737.  CLEARED.
Illegal block #7 (1691482175) in inode 6897737.  CLEARED.
Illegal block #8 (4257098933) in inode 6897737.  CLEARED.
Illegal block #9 (1728738978) in inode 6897737.  CLEARED.
Illegal block #10 (2320111554) in inode 6897737.  CLEARED.
Too many illegal blocks in inode 6897737.
Clear inode? yes

What's the best next step from here? Shall I run a short or long disk check? Or is my hard disk broken?
 
Personally I'd:
  • Run a long (off-line) disk check. That'll tell you if the HD is broken. If it is completely stuffed, well, it's hopefully still under warranty. If it isn't broken
  • Fix any pending sector reallocations, if there are any.
  • Copy all your files off.
  • Reformat
  • Copy them back on again
Others might be able to advise on the illegal block problem - but a reformat sidesteps the issue, if you've enough external storage to copy it all off.

EDIT: I've been here before myself - see https://hummy.tv/forum/threads/graded-stock-shipping-with-strange-filesystem-format.4793/

The conclusion of that thread was that the standard Humax firmware formats the disk in a way that fixdisk doesn't like. My solution was to format it form the command line, not using Humax's format. But I wasn't sure that was still an issue. Let us all know how you get on.
 
Last edited:
Thanks for the response.
I've done the long check and all is ok.

I've just checked and my pause/play delay issue has gone away, so I'm going to hang fire on the reformat for now.

I've acknowledged the disk fault, so will keep an eye out for more. If it does happen again, then I'll try the reformat as you suggest
 
Web interface version: 1.4.4-4
Custom firmware version: 3.10 (build 2734)
Humax Version: 1.03.12 (kernel HDR_CFW_3.10)
Loader Version: a7.30

I'm guessing you're going to suggest an upgrade?
 
...
EDIT: I've been here before myself - see https://hummy.tv/forum/threads/graded-stock-shipping-with-strange-filesystem-format.4793/

The conclusion of that thread was that the standard Humax firmware formats the disk in a way that fixdisk doesn't like. My solution was to format it from the command line, not using Humax's format. But I wasn't sure that was still an issue. Let us all know how you get on.
Indications are that the HumaxTV blob doesn't call mke2fs with some strange parameters but rather has embedded something like its source-code (obviously not actually that code, which is copyrighted and distributed under the GPL) calling out to /usr/lib/libext2fs.so.
Code:
# cat /etc/model
HDR
# strings /usr/bin/humaxtv | grep -E 'e2fs|mkfs|ext2|fsck|smartc'
libext2fs.so.2
ext2fs_default_journal_size
ext2fs_create_resize_inode
ext2fs_close
ext2fs_initialize
ext2fs_get_device_size
ext2fs_mkdir
ext2fs_allocate_tables
ext2fs_add_journal_inode
ext2fs_open2
ext2fs_write_new_inode
ext2fs_mark_generic_bitmap
ext2fs_read_inode
ext2fs_flush
ext2
e2fsck -F -j ext3 -p %s
e2fsck -F -j ext3 -f -y %s
smartctl --test short /dev/%
smartctl --log=selftest /dev/%s | grep '# 1  Short offline       Completed: read failure'
I don't think anyone has yet presented dumpe2fs -h ... output for a partition formatted by the Humax s/w versus one formatted manually?

Further observations:
  • The smartctl command templates are fairly certainly used for the Settings>System>Data Storage>HDD Test UI function, as you can see by running smartctl -c ... while the HDD Test is active.
  • The e2fsck command templates (the first for "preening", the second for forced repair) are strange, referencing some external journal file ./ext3. Under what conditions could they possibly be used or be useful?
 
Just to let you all know, I've upgraded to 3.13 and fixdisk still spins. It does finish eventually, but running again produces the same output. Lots of the following:

Code:
Clear inode? yes
Too many illegal blocks in inode 7511201.
Illegal block #10 (2320111554) in inode 7511201. CLEARED.
Illegal block #9 (1728738978) in inode 7511201. CLEARED.
Illegal block #8 (4257098933) in inode 7511201. CLEARED.
Illegal block #7 (1691482175) in inode 7511201. CLEARED.
Illegal block #6 (3914510362) in inode 7511201. CLEARED.
Illegal block #5 (4233493169) in inode 7511201. CLEARED.
Illegal block #4 (2962947696) in inode 7511201. CLEARED.
Illegal block #3 (820263210) in inode 7511201. CLEARED.
Illegal block #2 (4055449172inode 7511201. CLEARED.
Illegal block #1 (3184809161) in inode 7511201. CLEARED.
Illegal block #0 (2024937636) in inode 7511201. CLEARED.
Inode 7511201 has illegal block(s). Clear? yes

There are no reported errors against the disk though, so I don't currently have any problems, but figured I should let you know about fixdisk.
 
Just to let you all know, I've upgraded to 3.13 and fixdisk still spins. It does finish eventually, but running again produces the same output.
...
There are no reported errors against the disk though, so I don't currently have any problems, but figured I should let you know about fixdisk.

Are you saying that fix-disk runs successfully on partition 2 but then finds the same, or similar, filesystem errors if run again after rebooting into Maintenance Mode, and that no errors are reported on partitions 1 and 3?

If so, I think people might be interested to see the output of for dev in /mnt/hd[12]; do umount $dev; dumpe2fs -h $dev; done (run from Maintenance Mode).

Also, I hope someone has reported these "Illegal" blocks to CrimeStoppers.
 
There are no reported errors against the disk though, so I don't currently have any problems, but figured I should let you know about fixdisk.
That's in e2fsck territory, and nothing we can really do anything about.
The filesystem seems to be corrupted beyond any reasonable fix.
I think your only option now is backup, re-format, and restore.
 
Back
Top