Two Problems While Trialling CF 1.03.12 on HDR-Fox T2

How else???!! Observe where it says "Local Time is".
Did you perform a SMART scan? Where the results ok?
Did you perform a file system check? What where the results?
I was asked if I'd performed a SMART scan so was wondering how I initiated this explicitly - until I spotted the delay between clicking Disk Diagnostics and the Disk Information table, etc appearing. Hadn't spotted the timestamp.

Having read through https://hummy.tv/forum/threads/steps-for-repairing-a-disk-of-unknown-faults.3082/ I'll give fixdisk (aka "file system check"?; "disk tidy-up"?) a go tomorrow; hopefully the results report will be straightforward to interpret.
You need to poll for the results after the advertised time. They won't suddenly appear out of the ether somehow. This involves either looking at the data in the Web Interface again or learning how to use the smartctl utility from the command line.
How does a "short/long HDD test" differ from a SMART scan? I'd assumed (not unreasonably, I felt) that having issued the command the results would appear automatically in due course. Specifically, what data in the Web IF, please?
 
How does a "short/long HDD test" differ from a SMART scan?
SMART is a read-out of existing internal live performance statistics. The test runs are built-in, but require taking the HDD off-line. When you trigger them (via the maintenance mode menu), the HDD reports how long it will take (this is reported in the terminal session), and then all you have to do is reconnect the terminal session at (or after) the appointed time. IIRC, a short test is performed as part of fixdisk. All this would become apparent if you just got on with it.
 
What's the trick, please?
Oh, I was somewhat naive about that. It would be very nice if the Maintenance Mode "check" option returned the results. (Update – it does! For some reason it happened not to work when I tried it.)

The short test is run automatically as part of the fixdisk routine (long test is an option), and the results from those are parsed and output to the terminal session (which is virtualised with abduco and can be reconnected to as described previously).

The test results are logged internally and reported in the SMART output. This is available in WebIF as previously noted, but from a Telnet session you need to get to know the smartctl command:

First, identify the device designation for the HDD. This is probably /dev/sda, but that can change if you have a USB device connected at boot time. The "df" command will tell you which device is mounted as "/mnt/hd1" (if it's mounted at all – it doesn't have to be).

smartctl -a /dev/sda

...dumps the SMART stats to terminal. If that doesn't give sensible results, try "/dev/sdb" (etc).

Take note of the current Power_On_Hours figure, because that's how the HDD indexes its events log. (I don't know whether this is true of all HDDs, but the standard Seagate drives only store 16 bits of POH so the figure wraps at 65535 hours.)

At the end of the SMART dump you'll see the self-test results logged, with the most recent one at #1. You'll also see an in-service error log, which does not appear in the WebIF diagnostics.

Footnote: I have requested the SMART output is made available as a Maintenance Mode menu item and a terminal command.
 
Last edited:
...and use the "check" option:

Please select option: check

# 1 Short offline Completed without error 00% 32679
Hmm. That foxed me, because when I tried it out before editing my post 44 I got nothing out of it. I might have looked during a critical timing window between the test running (in which case "check" tells you how far it is through the test) and results being available.

Yes, that "32679" figure is definitely the POH.
 
you need to get to know the smartctl command:

First, identify the device designation for the HDD. This is probably /dev/sda, but that can change if you have a USB device connected at boot time. On the command line, type "df", then look for partition "/mnt/hd1" in the output. Use "/dev/sda", "/dev/sdb" (or whatever it might be) listed on that line, ignoring the partition number. Presuming the HDD is /dev/sda,

smartctl -a /dev/sda

...dumps the SMART stats to terminal.
Hmm... (No USB device connected.)

Please select option: cli

-----------------------------------------------------------------------

Humax HDR-Fox T2 (humax) 1.03.12/3.13

To return to the menu, type: exit

humax# df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/root 22144 22144 0 100% /
tmpfs 65536 136 65400 0% /tmp
tmpfs 62508 0 62508 0% /media
/dev/mtdblock1 2048 596 1452 29% /var/lib/humaxtv
/dev/mtdblock2 2048 528 1520 26% /var/lib/humaxtv_backup
humax#
 
I think that's because you're still in Maintenance Mode – the HDD isn't mounted as a file system. I'm not sure how to identify the correct device.

Just try smartctl -a /dev/sda.

(Post 44 updated)
 
Last edited:
Well that's odd, because even in Maintenance Mode I get this:

1700480721299.png

It's possible that our differing activity history (since last boot) has resulted in different mount statuses.

NB: I thought I had removed the busybox package, but looking again I see it's still present. I don't know whether that has any effect, but I doubt it.
 
Last edited:
Just try smartctl -a /dev/sda.
That worked ok and I've saved the SMART stats, etc as a .txt

I've run fixdisk twice now - I saved the second session's output to a .txt; it finished:

Finished
fix-disk: session terminated with exit status 0

Press return to continue:


Unsurprisingly the second run was a lot shorter than the first one. I had to take the dog out during the first and when I returned the laptop had gone to sleep and the session had disconnected. Not sure if that session's full output would have been available, but I don't have a .txt for the first run (which would have been more interesting, I guess).
 
..
Please select option: check

# 1 Short offline Completed without error 00% 32679
That's good. No SMART errors. So your drive should be physically fine.
Hopefully the maintenance fix disk procedure has performed it's magic and tidied up the filesystem.
When you reboot into the normal mode, you may see fix-disk files on the WebIF/diagnostics/View Log Files (on the right). They should be the logs of the fix disk runs. If they show changes were made, it should be for the better. Don't change anything. Use the HDR as normal for a few days to see if it made any difference.
 
Last edited:
I thought I had removed the busybox package, ...

As busybox is a single program that tries to implement the POSIX utilities* for low resource environments (looking at you, HD/HDR), it has endless build options to control which utilities are included and how fully each one is implemented.

The OEM firmware contains a (read-only) version of busybox, and so the CF also has one, which is what you get in Maintenance Mode (at least initially). In principle the CF version could be (and is, I believe) slightly different, with newer fixes and/or additional utilities (like stat, please), provided that not too much additional space is used and that no behaviour needed by the other Humax-supplied firmware components is broken.

There is a CF package of busybox with a build that adds more utilities and functionality (I posted a comparison ages ago). For example, the version of the POSIX shell (ash) supports the command ... utility, omitted from the firmware version. Other packages also provide implementations of various POSIX utilities, such as coreutils.

* Eg, when you run the command grep, whose path links to the busybox program, it sees the command name and runs its grep subsystem.
 
How does a "short/long HDD test" differ from a SMART scan? I'd assumed (not unreasonably, I felt) that having issued the command the results would appear automatically in due course. Specifically, what data in the Web IF, please?
I think a "SMART scan" is probably "alternative" terminology for starting either a short or a long test. The results appear in the output of "smartctl -l selftest /dev/sda" at a command prompt, and should appear at the bottom of the Disk Diagnostics page in the WebIf in a table titled "Self-test logs", but I just tried it on one of mine and the table was blank. This turns out to be a bug in the WebIf code for which there will be a fix shortly.
The test runs are built-in, but require taking the HDD off-line.
No they don't. You can run a "smartctl -t long /dev/sda" type of command with the unit operating normally, but you have to do it manually.
I think that's because you're still in Maintenance Mode – the HDD isn't mounted as a file system.
Mount status doesn't stop you accessing the device. How would you mount the filesystem if it did? In any case, here we use filesystems within partitions, so mounting the HDD is meaningless.
Not sure if that session's full output would have been available
Look in /mnt/hd3
In principle the CF version could be (and is, I believe) slightly different, with newer fixes and/or additional utilities (like stat, please)
That's been done in the beta build.
 
WebIF >> Diagnostics >> View Log Files >> fix-disk.#.log
Obliged.

Fix-disk.0 is the run that finished at 10:20 this morning and is 142 lines (already have this as a .txt)
Fix-disk.1 is the run that finished at 09:31 this morning and is 3,100 lines.

Is there an easy way of getting the latter into the same .txt format as I did with the former (attached)?
 

Attachments

  • 2nd fixdisk run.txt
    25.8 KB · Views: 1
Back
Top