I am sorry you have had such a bad experience with this, most disks seem to have just a few errors to fix.

Could you select 'diag' from the telnet menu, then enter 'diskattr'. If you could paste the results here, then we may get some clues as to the state of your disk.

Thanks, Martin, for your help again. And, XYZ, I will try and do what you say. Though not sure how to paste from a Telnet session. Hope it will be obvious. Guess I could take a photo of the screen?

I appreciate that we may just be unlucky and given the huge number of bad blocks found it may just be a sudden catastrophic failure of the disk (we had no prior warning or issues and it is 4 + years old).

Glad to hear that the disk fix programme does not usually require 10 hours continuous monitoring.

Will get onto your suggestion right away. Thanks.

Okay, I did that and all it put on screen was:

/bin/sh:/mod/bin/smartctl:not found


Whatever that means?

Have I done something wrong?

I don't think you have done anything wrong. It means the the script (custom firmware speak "diagnostic") couldn't find a program it needed (smartctl) which reports the health (or otherwise) of the disk. I suggest wait for xyz321 to respond.

Thanks, Martin. Will do.

I guess I have no real options left then but to try to reformat the hdd and hope that actually makes it recordable again.

If not then it presumably means that the box is not salvageable.

Might try one last run of disk fix today. I assume I can just abort it by switching the box off if it starts taking forever again.

I am hoping all the bad blocks it said it fixed on Friday really were fixed so that should speed it up???

But it can hardly make things worse.

See what happens and then try the formatting tomorrow otherwise.

It is already finding bad blocks and reverting the countdown, so my hope that it would tick through to the end and start searching for inodes is unfounded.

I assume that it did not really fix the bad blocks it said it did on Friday. Unless these are all new ones?

It has found 5 so far and the countdown is restarting around 3856 seconds. That is a little better than Friday. So maybe....

Though it has reached that point an hour faster than Friday it then took another 6 hours to get from there to about 150 seconds lower!!! So if it only fixed the early blocks and has all the later ones to do again that will just not be doable.

Will persist for a bit but no way am I going through another 9 hours of constantly pressing Y. So if this keeps happening again I will just press x to see if that cancels fix disk, though I suspect it will not as it did not on Friday.

If not I will just quit by rebooting the box.

Been running the test now for about 30 mins and it is at 3832 seconds. It is about 3 - 4 hours ahead of when it got there on Friday. But it is still finding bad blocks. 11 so far. Though that is easily 100+ fewer than at this stage Friday.

Will keep going a little, but if it does what it did on Friday then it might still find another 150 or so and take about 4 more hours.

Will keep persisting as long as possible.

I am sure this 'exciting' thread will go down in DS history.

However, someone using this utility might find it helpful one day.

Okay, 1 hours 15 mins into the repair and still finding bad blocks but still less often than Friday with larger gaps in the LBA number.

24 so far and the countdown is at 3766 seconds which it reached Friday after around 5 hours and many times that number of Y presses.

One hours 45 minutes in, 34 blocks fixed, countdown at 3680.

Sorry I have been very busy recently (worked 33 hours in the last 60 hours). Based on what you are reporting the disk is in a terrible state and needs replacing; a new hard drive of the correct type costs about £53 for a 1TB model.

I am sorry your time has been wasted. In many cases a single pass of fix-disk will correct a small number of errors and render the disk fit for service. Every failure is different and without trying you can't know whether a disk can be saved.

I would not be able to replace the HDD so would just have to buy a new box. Do you think this means a reformat will not make it work either?

It did finish the disk fix very quickly (3 hours) and found just 39 bad blocks this time. Many, many less than before.

It is now on the interminable bit at the end where it says 'searching for inode' - whatever that means. It did this for over half an hour just paused on screen for the first of the 39 bad blocks. Then it put some text up about 'can't read next inode whilst doing inode scan'. Then started looking at the next of the 39 blocks and says something about a block being marked as in use.

It has been 'searching for inode' on this one for ages now.

I have no clue what all of this means or if it is good or bad. But it seems to require me not to be watching it so I am just leaving it be to go on its search, checking every now and then.

If it takes this long over all 39 blocks it supposedly repaired it will take many hours to finish I guess.

Lets see. But is not the fact that this run was faster and found far less issues than the first run on Friday good(ish) news?

Let the fix-disk run to completion. The format now that fix-disk has cleared up a lot of problems might help temporarily but my best guess is that the hard drive is end of life. If you wish, I will fit a new hard drive for the costs of the parts and the postage as I feel bad about wasting your time. If you can get the box running for long enough to install the full web interface for the custom software then you can look at the diagnostic data for the hard drive which will confirm whether or not the disk is beyond hope.

Please in no way feel bad about anything. You have been truly fabulous and we are really grateful for all you have done.

It has been an experience. So no need to make your very kind offer.

I will try to let it run to completion. It is currently searching for inodes on the 4th of the 39 fixed blocks. But not sure it actually finished the other three as it just seems to try then moves on after a long time of searching to the next. No text suggesting it achieved anything on the last one. Just that the block assigned was 'in use'.

Will keep reporting what does/does not happen but, please, do not at all feel the need to keep watching my posts. Check them whenever you have a moment tomorrow.

Have a rest after your long working weekend.

And we are really very grateful for all of your support.

It just moved onto the 5th block of 39 after puttng up the same message as after the other 4 - 'Cant read next inode whilst doing inode scan'. Whatever that means.

PS: What is the full web interface software?

I think this inode search for the remaining 34 blocks will literally go on to breakfast time. Hope my computer will let it do so whilst I am asleep. Even in suspended mode. And that we have no power cut!

What you are currently running via telnet is a minimal version of the Custom Firmware. The full version allows you to point a web browser at the Humax and opens a whole world of possibilities. Personally we don't use many of them but the ability to monitor the health of the hard drive and to restore the schedule of future recordings after a retune are very helpful. However to install the full version you need to have a working hard drive which obviously you currently don't have. I am of the opinion that you shouldn't waste more time with fix-disk. Try formatting and see if that succeeds.

It is one o clock, as you can see. Fix disk has been running forever now (well 11 hours in total).

It WAS checking inodes very slowly. I left it running by my bed and after 6 hours of inode checking it was up to the 14th repaired block of the 39. Though I am not sure it did anything with any of them.

I fell asleep and rechecked and it was still doing number 14 90 minutes later. It never stayed on any inode that long and was totally frozen.

I could not get any response from the programme and a message said 'network error'.

When I tried to open Putty I was told it was not safe to do so.

I have had enough tonight, and, I believe,that's it for me now.

I cannot attempt this again. And any programme MORE complicated than this one will be way beyond me, I am afraid.

Am going back to sleep as it is a busy day tomorrow with nurses, doctors and carers.

Thanks for staying with me.

Okay, I had to do something to terminate the session and after doing so decided to take your advice. The box still had the format hdd message on it. I decided enough was enough and to try that.

It appears to have worked, in that it shows space and so on.

I have set a couple of programmes to record overnight to test it.

Will see tomorrow if they are OK. The red record light DID come on to start the first.

I know the recordings are gone and after more than 24 hours of disk fix that is a wrench...but I could not have done any more.

Hopefully the repairs have in some way made the empty disk serviceable.

Maybe we can discover that tomorrow somehow via the test you mention.

Have a good night (what is left of it).

Okay, well so far, so good. The things I recorded last night played properly. The ability to go back a few hours on the channel you are tuned to is there (which needs the hard disk to work, of course) seemingly operating as it used to do. Oddly the minor lip synch errors I always got before have even been cured now and sound matches picture.

However, when I tried the diskattr test again it still gave the same result as it did when nothing was working.

Not sure that means anything about the state of the HDD.

With the Humax connected to the router and not in standby, put the IP address of the Humax (which you have been using for telent) into the address bar of the browser. This should bring up a page with a button inviting you to install the full web interface of the Custom Firmware. Press the button; it will take a few minutes to download and install the interface. When it is installed click on Diagnostics and then click on Hard drive. Tell us the raw value (first column) of the following attributes: ID 5, Reallocated_Sector_Ct, ID 9 Power_On_Hours, ID 197 Current_Pending_Sector, ID 198 Offline_Uncorrectable. Which attributes are labelled in the When Failed column?

Ah, that is odd because the page says that to do this the Humax needs a hard disk and none has been found.

???

There are three partitions on the hard drive; the first one holds the cached EPG data and the DLNA databae (and where the Custom Firmware also holds data), the seconds holds the actual recordings and the third holds the buffers used by streaming content such as iPlayer. My best guess is that the first partition is unusable; a clue to this would be the program guide being much slower to fill with data.

I still think the drive is well beyond its end of life and you will in all probability lose another collection of recordings before long.

Thanks. I guess that makes sense. Would that be why fix disk took 8 hours looking for inodes and seeming to find none of the 14 searched in that time?

The EPG seems fine. Runs the full week ahead. Just did a search on it and it instantly revealed 35 items on the search word.

Tried i player in HD and it took a while to start playing, but I was expecting that. Never used it on the Humax before but it played fine within a minute or so.

Tried more recordings this morning and they are fine and show no obvious glitches.

At present it is working, which is a bonus. I like the box, and, having had a quick look at views about what Humax have just replaced it with opinion seems to be not as good as this box. So would not be in a hurry to swap for one of those unless I am missing something like faster, quicker, more reliable, better features etc.

Having 427 Gb (?) free space in the HDD might help. Will try not to keep loads of recordings there in future.

No doubt the drive is on its way out as you say. Then I will have no option but to replace it. That might be tomorrow, of course. We shall see.

But thank you for all your assistance and I will get your USB stick back in the post on Wednesday when Crossroads come to look after mum and I can get out.

inodes are the data blocks that the file system uses to keep track of where things are. There is some redundancy but in your case it was hopeless. I don't think the failure to find inodes is directly linked to the problem with the first partition but just generally reflects the poor state of the disk.


I don't think you read what I wrote. The EPG is initially populated at startup from the Data on the hard drive and is then refreshed using data transmitted over the air. The indication that the first partition isn't available would be that instead of the EPG being fully populated within a few seconds of switch on it would take several minutes to fill.

The best way forward would be to have the hard drive replaced but you seem to be discounting that option which is your choice.