Missing (virtual) remote, with added disk woe.

One of our HDR;s now has 27,000 hours on the clock, and for a long time now it has been sending me daily emails, the content of which is

Device: HDR-Fox T2 (HumaxK - 00:03:78:ba:32:97)
At least one failed recording has been detected in your
media library. The following recording(s) failed:

/How to Beat___ Stress/New_ How to Beat___ Pain_20200326_2100


- Aha - I've just noticed the programme title. - How to beat stress - by not watching the TV and wrestling with all the tech in the house (there's way, way more than just HDR's.)

anyway, the point of this is that an internal check must be looking at the VTOC and the data, and deciding that a) a recording failed, I guess because the end of extent wasn't written, but then, I can find no entry by this name amongst the recorded items, failed or otherwise.
So time went by
until I ran a full disk check, with no result. Nothing reported, and still no visibility of this item.

a bit more time has gone by - a week or so.
and I found that when I log in from an iPad that the virtual remote has vanished. If I attempt to browse to the page by name, I get 'no page by that name'

And all of that brings us up to date.
This a.m. I ran the RMA option, then, according to Black Hole's note, cancelled from telnet, and subsequently attempted to reload from USB, using the same device and content as used a month or two back on the other Humax.
This time, it stalls on the 'Humax' splash screen. Nothing to do but use red switch cancel and an unplanned power down.

And there we have it. I tried once more to load software, same result. The humax will start normally, if no USB present, the WebIF loads, but the Virtual remote is missing. The added woes come from the discovery that attempting to use disk diagnostics produces an error message (attached)

The only package now showing installed is the WebIF, and normal functions are normal. I'm hoping to avoid doing a full format.

My thoughts are to run another fix disk.
When I telnet to the box, and request maintenance restart, and reboot, and telnet back in, I now get a message

"note" some areas of the hard disk are not mounted. Not all functions will be available. (Normal if you have just run a disk check)"

Well, unwelcome, but explanatory, and adds weight to the original conjecture, that an unreported disk issue is lurking.

How about running fix web?
tried that. The output is:-


Please select option: fixweb
killall: lighttpd: no process killed
2020-08-26 11:14:01: (network.c.309) can't bind to socket: 0.0.0.0:80 Address already in use
Removing package webif from root...
Not deleting modified conffile /mod/webif/html/favicon.ico.
Installing webif (1.4.8-8) to root...
Downloading http://hpkg.tv/hdrfoxt2/base/webif_1.4.8-8_mipsel.opk.
Configuring webif.
/mod/bin/jimsh: can't load library 'libcrypto.so.1.1'
/mod/bin/jimsh: can't load library 'libcrypto.so.1.1'
Collected errors:
* file_md5sum_alloc: Failed to open file /mod/webif/html/favicon.ico: No such file or directory.



And that's where we are.
Am I right in thinking that this is looking like a new disk?

Or do I begin at the logical start of fault clearing again, and stay focussed until a full disk check and re-instal has executed successfully ?


At least I haven't lost any smoke, and it's all inside still.


RScreenshot 2020-08-26 at 11.02.34.png
 

Black Hole

May contain traces of nut
This a.m. I ran the RMA option, then, according to Black Hole's note, cancelled from telnet, and subsequently attempted to reload from USB
Setting RMA and then clearing the flag should not require re-installing the CF as a firmware update. Please ensure you have followed the procedure exactly as per the instructions you mentioned:
Reinstate Corrupted CF via RMA Mode

The following will (should) wipe the CF from the HDD so that it can be re-installed from scratch. It is not necessary to re-install the custom firmware to Flash by firmware update.
  1. Establish a Telnet connection (for details see https://wiki.hummy.tv/wiki/Telnet). If asked for your PIN, the default is "0000" (this can be changed by the user). If not presented with a menu, at the command prompt type "tmenu".
  2. From the Telnet menu select option "rma".
  3. Reboot.
  4. Re-establish a Telnet connection (as above), and from the menu select to option to reset RMA mode.
  5. Access the HDR-FOX's IP address from a web browser – the option should be available to re-install CF.
 
Thanks.

I had first run the 'download and install the Full Web Interface' option. This did not result in the remote control option appearing, so, knowing that the usb stick has a known working version, I reloaded from that. All of which didn't affect anything.

I have just now repeated the 'Prepare RMA - cancel - Download and install' sequence, which reports successful conclusion


on rebooting, I now have a red warning saying 'potential hardware problems with the HDD. Overall health assessment is : Unknown. Go to disk diagnostics.

Still
SMART status is 'passed' Though still no remote control option on the webIF
All of the foregoing, and this message, seems to me to be saying that the disk is dying. I seem to have a couple of 500Gb drives looking for work, but maybe a shiny new 2Tb would work these days? - have to go and read the threads about large disks.

Attributes
IDNameFlagsRaw ValueValueWorstThresholdLife LeftNotes
1Raw_Read_Error_RatePOSR--135303978117099006-
3Spin_Up_TimePO----0097097000-
4Start_Stop_Count-O--CK2421007707702072%-
5Reallocated_Sector_CtPO--CK0100100036100%-
7Seek_Error_RatePOSR--498579288087060030-
9Power_On_Hours-O--CK2776506906900069%-
10Spin_Retry_CountPO--C-0100100097100%-
12Power_Cycle_Count-O--CK1210508908902087%-
184End-to-End_Error-O--CK0100100099-
187Reported_Uncorrect-O--CK0100100000-
188Command_Timeout-O--CK0100100000-
189High_Fly_Writes-O-RCK1099099000-
190Airflow_Temperature_Cel-O---K55045 (55°C)040 (60°C)045 (55°C)FAILING_NOW
194Temperature_Celsius-O---K55055060000-
195Hardware_ECC_Recovered-O-RC-135303978042023000-
197Current_Pending_Sector-O--C-0100100000-
198Offline_Uncorrectable----C-0100100000-
199UDMA_CRC_Error_Count-OSRCK0200200000-
Self-test logs
No.DescriptionStatusRemainingWhenFirst Error LBA
# 1Short offlineInterrupted (host reset)00%27579-
# 2Short offlineCompleted without error00%27579-
# 3Short offlineCompleted without error00%19872-
# 4Short offlineCompleted without error00%19640-
# 5Short offlineCompleted without error00%10300-
# 6Short offlineCompleted without error00%9181-
# 7Short offlineCompleted without error00%9180-
# 8Short offlineCompleted without error00%7349-
# 9Short offlineCompleted without error00%4807-

Acknowledge any current disk faults.
 

/df

Well-Known Member
First off, this is a Customised Firmware issue, which a kind Moderator might wish to relocate.

The problem OP saw is not unprecedented.

However it shouldn't have happened if the BH procedure was followed and worked correctly.

Nothing to worry about on the disk stats, but people will probably advise you to install the fan package, and they could be right. You can trade more fan noise for a potentially longer disk lifetime.
 
I looked at the referenced topic. I don't think it is directly relevant. The remote control went missing at some unspecified point. I can point to no actions that may be responsible.
I find tat as of now, however, that 'browse files' over the web is showing only a portion of the VTOC - as far as the letter Q.
That suggests again to me that the disk has sustained some damage. Therefore back to disk diagnostics and scans.


The over temperature is a first. At a point, the fan utility was installed, but right now there is only the most basic IF in place.

Thanks for the suggestions. I'll report in with any further information.

R
 

prpr

Well-Known Member
That suggests again to me that the disk has sustained some damage. Therefore back to disk diagnostics and scans.
There's nothing physically wrong with your disk. If fix-disk from maintenance mode runs clean, then any further problems are elsewhere.
Run the fix-flash-packages diagnostic to restore the ir remote package, once you have fixed/checked lower level things.
 

/df

Well-Known Member
When people run fix-disk I suppose they expect it to fix disk things without necessarily understanding which device is involved. Perhaps it should also run, or offer to, fix-flash-packages?
 

prpr

Well-Known Member
When people run fix-disk I suppose they expect it to fix disk things without necessarily understanding which device is involved. Perhaps it should also run, or offer to, fix-flash-packages?
Sounds like the tail wagging the dog to me.
Should it also go round the disk looking for orphaned .hmt files?
Or whatever other fault conditions somebody could dream up?
 

Black Hole

May contain traces of nut
fix-flash-packages, as mentioned in other posts above, could have been your missing link - I presumed ir wasn't there because you hadn't installed it, they presumed you had installed it but it wasn't working.
 
When people run fix-disk I suppose they expect it to fix disk things without necessarily understanding which device is involved. Perhaps it should also run, or offer to, fix-flash-packages?

Well, I wouldn't.
The solution, as to so much in life, is to provide enough information to enable someone not completely conversant with the underlying OS and package architecture, to evaluate what needs to happen.
This starts with adequate explanation within an error message. I have spent a lifetime pointing out that en error message saying 'there is an error' is completely useless, merely sowing uncertainty and FUD in the mind of the innocent user. Whereas an error message saying 'Function XYZ reports unable to write end of filmmaker, Error code 7654' is far more informative, and a detailed explanation can be found in the documentation of error 7654.

Fixdisk, is a misnomer, it is not fixing the physical disk. All it can do is identify, and perhaps correct, recording errors, maybe at the cost of further damage to files. For the public to work this out, you have to tell them. Most of 'the public' have no idea how computing works, or even how computing could and should work, having been brought up with windows.

Even many 'technicians' these days are so insulated from the bits and bytes that they have little understanding of platter level operations. (We ARE talking spinning disks still in the Humax under discussion)

So running fix flash as well wouldn't help, unless the problem is understood. By all means , provide a 'diagnostic' as a package. I refuse to use the AI terminology, for it has recently been so abused, but a modest Knowledge based dialogue could ripple down symptoms and potential problems to localise the problems, and then suggest remedial actions, offering to run them if allowed.


Take the case I presented - it comes in 2 parts.
1) The cause of the scary email messages. - the text is embedded in code somewhere, so we could instantly localise what produces it, and the precise reason for that message being edited (if the same message is used as output for multiple tests - see above, and add a unique identifier for each emitting true/false comparison test.)
From knowing what condition caused it, we can expand on the reason, and suggest corrective action. Another response to my query here says that it is an orphaned file that can be disregarded as it takes up negligible space. All well and good in the land of windows technicians, but in my world we don't leave ever more clutter on the disk(s) and we certainly don't provide an error with no solution. If we know what it is, and can describe it, and the system can find it, then we can damn well provide a tool which will remove it. A switch for whatever found it (that uniqe I'd will tell us what code to change) that informs the function to delete the offending object. After running in find and delete mode, the function will reset itself to find only.
That would clear all that up.

2) the missing remote function.
Symptom - missing remote.
Solution - install ir package. Do you want to install it now?
and that would clear all that up.


I'm happy enough to dig about with the CF - it's an excellent extension of the box, and does all I need and much more. I'm very very grateful to everyone who developed it, and I know the picture I paint is not in scope for such an enterprise, so I tinker on this strange box, running under an unfamiliar OS, with tools I don't fully understand, with minimal documentation. Because I just love the box. My daughter moved out and bought a much newer, larger, Humax, and it's all shiny and rounded, and the software seems fixated on expensive streaming services and portals and suchlike. Her vision issues mean that the busy screens are much harder for her to see and navigate than the HDR-FOX. Maybe that's why she's staying here while the plague passes, and filling my boxes with stuff she wants to keep on her PC.


If this comes across as a bit rants - it really isn't. Regard it as Yorkshire bluntness.

R
 
There's nothing physically wrong with your disk. If fix-disk from maintenance mode runs clean, then any further problems are elsewhere.
Run the fix-flash-packages diagnostic to restore the ir remote package, once you have fixed/checked lower level things.
Sorry - I didn't intend to suggest physical damage. 'damage merely indicates that recorded material is no longer accessible. All it takes is a bit missing - or added.

As it happened, fix disk ran with dozens, hundreds of changes.
 
fix-flash-packages, as mentioned in other posts above, could have been your missing link - I presumed ir wasn't there because you hadn't installed it, they presumed you had installed it but it wasn't working.

I set about returning to a valid, known state, and it hadn't occurred to me that that small part of the WebIF would need installing separately. My own fault for not sitting down to read the full documentation on the package content. - And me the only one who subscribes to RTFM in the household. !
 

Black Hole

May contain traces of nut
As it happened, fix disk ran with dozens, hundreds of changes.
These "changes" are correcting where the file system has become disjointed. A file system is a database of indexes and pointers with redundancy, so a few broken bits don't necessarily kill it completely, but will benefit from a spring clean occasionally. Sometimes the indexing required to access the data for a particular file is unrecoverable, in which case that file is lost (but would have been anyway).

Modern file systems in combination with modern HDD architectures are able to resist a few soft or hard errors, until the number of errors starts to mount up as the HDD heads towards end-of-life - then there tends to be a cascade, with every attempt to repair the database introducing yet more errors.

If you've got away with it this time, fine... for the moment.
 
These "changes" are correcting where the file system has become disjointed. A file system is a database of indexes and pointers with redundancy, so a few broken bits don't necessarily kill it completely, but will benefit from a spring clean occasionally. Sometimes the indexing required to access the data for a particular file is unrecoverable, in which case that file is lost (but would have been anyway).

Modern file systems in combination with modern HDD architectures are able to resist a few soft or hard errors, until the number of errors starts to mount up as the HDD heads towards end-of-life - then there tends to be a cascade, with every attempt to repair the database introducing yet more errors.

If you've got away with it this time, fine... for the moment.

I would be mildly annoyed if it died unexpectedly, but it's a minor part of life in the bigger picture. A bigger disk would be the way to go, since reading here while ago that 2TB can now be supported.
 

Black Hole

May contain traces of nut
since reading here while ago that 2TB can now be supported.
You are not limited to 2TiB. Things Every... (click) section 12:
As of March 2016, the custom firmware* for HD-FOX and HDR-FOX includes facilities to enable installation and formatting of GPT disks. This means there is now no practical limit to the size of disk, if the user is prepared to go the custom route (returning to standard Humax firmware while a GPT disk is installed could result in erasure of the disk contents).
 

/df

Well-Known Member
To summarise what fix-disk does:
  • the tricky part, where it lives up to its name, tries to find any bad hard disk sector and return it to life by forcing it to be remapped (losing its contents) or declared as good (saving the contents), while keeping track of whether a remapped sector is used in any file or directory; it uses the disk's built-in SMART test to do this much faster than the traditional badblocks program (days->hours);
  • less tricky but still challenging for a large partition with the limited RAM available, it runs the filesystem check and repair e2fsck.
The Humax settop software seems to have a limited boot-time filesystem check but it can't run the full filesystem check and repair that fix-disk does, because of memory limitations. So filesystem issues can accrue over time, and the Humax software only offers reformatting as a solution. fix-disk makes it possible to revive a disk and its filesystems without removing the disk and plugging into a more capable system.
 
Top