• The forum software that supports hummy.tv has been upgraded to XenForo 2.3!

    Please bear with us as we continue to tweak things, and feel free to post any questions, issues or suggestions in the upgrade thread.

Health of HDD after fix-disk

confabulator said:
I've had some more offline uncorrectables develop and am now at 232.
Click to expand...
Decline seems to be accelerating. Not good at all.
What I'd like to do is immediately install a new drive and put the old drive in my caddy to copy across to the new, but that has to wait until the new drive arrives. For now I'd like to back the recordings on the drive as I've got several 1TB externals going spare here.
Click to expand...
I would take out the suspect drive out and put one of these 1TB drives in instead, until you get the new drive. You don't want to be writing to the faulty one at all now. The time-shift buffer is constantly writing, even if you're not recording anything.
Would you suggest running Fix-disk as well as writing off the drive?
Click to expand...
I don't quite understand what you mean by that. I would suggest NOT running fix-disk now.
I'm geting quite nervous to do anything with it to be honest.
Click to expand...
Yes, I would be too.
Should I have rerun fix-disk to try and recover these?
Click to expand...
The data is already gone by that point, so it wouldn't have helped.
 
Hi again,

So despite best intentions life took over and I've only just replaced the HDD today. I'm copying files off the old HDD as we speak. I've got two final questions please.

Firstly on the original HDD, now externally mounted, there are some vid files in "lost and found" directory. Can someone explain why? Some of these look like files I deleted but some look like they may be files that I lost during the disk corruption. Is that possible?

Second is there a way to run fix-disk on the old drive externally? I'd like to do it just to see what might recover, especially as I hit delete loop this week and I believe it's possible the files not apparently recorded while in this state may exist but not visibly. I could drop the old drive back in again and run fix-disk but I'd rather not fiddle with that fan cable again!

Thanks again.
 
confabulator said:
...
Firstly on the original HDD, now externally mounted, there are some vid files in "lost and found" directory. Can someone explain why? Some of these look like files I deleted but some look like they may be files that I lost during the disk corruption. Is that possible?
...
Click to expand...
lost+found is where the file system repair puts orphaned files and directories (eg, whose parent directory was partially corrupt or unreadable). So, yes, you might well be able to recover lost files from there (hence the name!).

...
Second is there a way to run fix-disk on the old drive externally? I'd like to do it just to see what might recover, especially as I hit delete loop this week and I believe it's possible the files not apparently recorded while in this state may exist but not visibly. I could drop the old drive back in again and run fix-disk but I'd rather not fiddle with that fan cable again!
...
Click to expand...
The actual fix-disk script is tied to the disk in a HDR, like this:
  1. find the disk
  2. find the partitions, repairing if necessary/possible
  3. repeatedly: run a SMART test until a bad sector is found, then force the remapping of the bad sector with hdparm
  4. check/repair the file system on each partition
You might have success by replacing the for loop some 190 lines from the end
Code: 
...
# If a disk appears to be on a "PCI bus" then assume that is the               
 # internal disk to be checked.                                                 
                                                                                
 for dir in /sys/block/sd? ; do                                                 
...
done
with dev=sdx, where sdx is the disk device corresponding to your old disk connected via USB.

Alternatively you could run the component tools on their own from the command line but a modern disk with even a small fraction of bad sectors will need an automated approach. Probably at this point you will get most value from just running fsck again.
 
/df said:
The actual fix-disk script is tied to the disk in a HDR, like this:
  1. find the disk
  2. find the partitions, repairing if necessary/possible
  3. repeatedly: run a SMART test until a bad sector is found, then force the remapping of the bad sector with hdparm
  4. check/repair the file system on each partition
You might have success by replacing the for loop some 190 lines from the end
Code: 
...
# If a disk appears to be on a "PCI bus" then assume that is the             
# internal disk to be checked.                                               
                                                                              
for dir in /sys/block/sd? ; do                                               
...
done
with dev=sdx, where sdx is the disk device corresponding to your old disk connected via USB.

Alternatively you could run the component tools on their own from the command line but a modern disk with even a small fraction of bad sectors will need an automated approach. Probably at this point you will get most value from just running fsck again.
Click to expand...

Thanks for the reply. I'll see what is available in lost and found and copy as desired.

I'm lost with the rest of that I'm afraid. Where do you even find the coding that controls fix-disk? If I try and follow that approach would that be as guaranteed as dropping the old HDD back in the box?
 
confabulator said:
If I try and follow that approach would that be as guaranteed as dropping the old HDD back in the box?
Click to expand...
The USB interface would slow everything down, even if you managed to patch the relevant files. Surely it's easier simply to connect the relevant disk in place of the internal HDD, even if you don't actually fit it physically?
 
Black Hole said:
The USB interface would slow everything down, even if you managed to patch the relevant files. Surely it's easier simply to connect the relevant disk in place of the internal HDD, even if you don't actually fit it physically?
Click to expand...
Thanks BH. I think that answers it. I don't know that I can connect the old HDD without replacing it physically but I'll see if the cables can swap over and make it easier.
 
Sorry to have to ask another question. I'm still in process of copying across the old vid files to the new HDD - as you know this takes time. I just went to review to recordings on web-if to discover I needed to reinstall so did that. The problem I have then run into is the new HDD didn't fully shut down to allow a reboot and I had to manually do this by using the switch at the back. The web-if is now up and running but I'm concerned about the lack HDD shut down. Any idea why? The first time the external drive was still connected by USB but I tried again after pulling it with the same result. I've change no settings on the Humax, there is nothing scheduled until tomorrow and I don't use auto on or off.
 
Thanks BH. I'd actually already looked at a thread based on that. It was clear earlier the unit was staying half-awake and failing to reach standby. There is no obvious reason from the set up. Could it be a fault with the new HDD? Is it likely I'd be able to copy over a file from the old HDD that is corrupted? Any idea if attaching and external HDD could make a difference. Any way to tell if I have a corrupted file other than deleting things and starting again? Sorry to ask a lot but frustrated.
 
That's a laugh! I never turn off my units and they are still going strong nine years later. Turning drives on and off is worse.

However, if you are determined to figure out which is the rogue file, what I have done in the past is a process of divide-and-conquer: First, back up half the recordings somewhere else and delete them. If the problem goes away, the faulty file is in the set you deleted, so restore half of them... and so on.
 
Thanks. I take your point which is reassuring.

I'd already thought about that kind of method too, especially as these have been restored from backup already, but wondered if there was a more cultured way. I'll probably delete folder by folder until I find an answer and if nothing clears it I'll re-format.

Unless someone else has another way?
 
One way you can sometimes identify rogue files where the Humax is stuck trying to process them is to to issue in a command window
Code: 
lsof -Fns | grep "\.ts" | sort -u
This lists inuse .ts files,
When the system is idle there shouldn't be any in use (apart from perhaps the TSR recording n/mnt/hd2/Tsr/0.ts)
If something is in use when system should be going to standby try renaming it to a different file extension, rebooting and trying to shutdown again
 
So I've worked out how to directly enter at a command prompt which was nice and simple. I wasn't sure but copied that code exactly as typed by Mymsman and got nothing back, just back to the command prompt. I then tried the lsof -Fns on its own and got a very long report. Was I supposed to enter that code exactly? If so there was no list produced.

For info the web-if says system is in standby but the disk is still running and web-if fully functional so not truly in standby.
 
confabulator said:
I'll try and work out how to use a command window later.
Click to expand...
Either install webshell and access via WebIF >> Diagnostics, or use a Telnet client on your PC.

Index of Package Primary Topics

[work in progress] If you are interested in the discussion around the introduction of a specific Custom Firmware installable package, a forum search for the package name enclosed in square brackets ("[...]") and restricted to the topic title (see the forum search box at top right) should find...
hummy.tv hummy.tv
​
wiki.hummy.tv

Telnet

wiki.hummy.tv wiki.hummy.tv
 
You must log in or register to reply here.
Back
Top