WebIF Failed

Black Hole

May contain traces of nut
I seem to have total failure of the WebIF on my HDR4. It normally just sits there recording and decrypting radio programmes and then converting them to MP3 for later consumption in the car, accessed by SMB but also occasionally by WebIF. The other day I noticed the RS service hadn't "seen" HDR4 for over a year, so I set about investigating.

I fired up WebIF Diagnostics and ran fix-flash-packages for good measure, then rs/sync, to no apparent effect. rs/checkin claimed there was no service. Things went downhill from there: WebIF Package Management claimed there were no packages installed, nor any available. Then the main WebIF page worked but wouldn't open any sub-pages. Then all I got from the WebIF IP address was a directory.

I've tried rebooting, but now I get the reinstall page. So far as I can see, the HDR itself is functioning normally. Obs I had better run fixdisk, and thanks to webshell I can do that on my iPad...
 
Holy mother of god. I don't seem to be able to copy and paste using webshell on the iPad, so here's some highlights:

E0B962FD-5447-4B55-8CC0-55E57A0B3291.jpeg
...

A4E0AB64-1337-480B-AA13-D4FA63237C33.jpeg
 
Progress appears to have stalled "cloning multiply-claimed blocks"... is it time to pull the plug?
 
I would say yes. (Is there anything of note on the SMART diags?)
As the drive may be iffy, I'd do this to try to reclaim the drive
  • remove drive - plug into a PC running Linux on a SATA3 interface (it'll be faster and you'll get better logs etc)
  • perform SMART diagnostic, check logs and dmesg (first time)
 
Last edited:
The figures listed by smartctl -a /dev/sda (run in parallel with fixdisk remaining active) don't look terminal. I think this is a major file system corruption rather than anything else.
 
I was being hasty – there has been further progress in the current fixdisk session.
 
Anyone know why fixdisk takes so long to "clone multiply-claimed blocks"? Still going (there are a lot of blocks to do, but the progress for each one is glacial).

It would be great to have some kind of "I'm alive" indication instead of an unchanging terminal session for hours on end.
 
Anyone know why fixdisk takes so long to "clone multiply-claimed blocks"? Still going (there are a lot of blocks to do, but the progress for each one is glacial).

It would be great to have some kind of "I'm alive" indication instead of an unchanging terminal session for hours on end.
Unless the underlying function is coded to return some periodic progress indicator it is very difficult for any script invoking that function to return anything meaningful
 
I would have abandoned it by now and connected it to my Linux PC and done it there. It will be magnitudes quick
But that would assume I know what I'm doing, whereas fixdisk just does it without me needing to dismantle anything (or even being present). It's no real skin off my nose to wait, so long as it's still making progress and not stalled.
 
Unless the underlying function is coded to return some periodic progress indicator it is very difficult for any script invoking that function to return anything meaningful
Well, it's e2fsck and it can report progress to an open file descriptor, say 3, with the -C3 option. Actually, fix-disk uses that and tees the output to a progress file and a log file. A background shell process reads the progress file and drives a spinner and %complete display in the VFD.
 
I don't know when exactly, but the session completed some time in the night. So that's between 52 and 60 hours. Judging by all the cloning of multiply-claimed blocks, there will be a significant number of recordings corrupted.
 
I don't know when exactly, but the session completed some time in the night. So that's between 52 and 60 hours. Judging by all the cloning of multiply-claimed blocks, there will be a significant number of recordings corrupted.
Speaking as a bystander in this conversation - not being a Linux expert or a webif user - maybe you should attempt to recover the recordings to another disk and then reformat the problem disk. In fact, if I knew what I was talking about I would have suggested this at or around post #4. It probably would have made sense to follow bottletop's suggestions. From my experience of other computer systems, it smells like a corrupt filestore. (How that happened I don't know - power cuts in the middle of some significant file operation used to do it for me). Hopefully your disk may be perfectly okay and provide years more service. If not, at least you know now and can replace it. Unfortunately you may have lost some recordings - unless you record "important" items on more than one machine. (I often do, but that's usually tv not radio and HiDef on one and StdDef on the other)
 
Well, I've just been trawling to find out why undelete stopped working last October and is no longer installed, and now I know why. Many sleeps since then.

For information, I've had no further problems. I'm running a fixdisk now...
 
Famous last words! I'm getting a raft of multiply-claimed blocks again. I think I'll abort it for the moment.
 
Sheesh. Now I can't connect in Maintenance Mode (wifi). This morning it was fine, but tonght I've rebooted several times and I can't access the MM menu, despite wireless-helper being installed.
 
Oh sh*t, no I can't - HDR4 is on WiFi!

Sheesh. Now I can't connect in Maintenance Mode (wifi). This morning it was fine, but tonght I've rebooted several times and I can't access the MM menu, despite wireless-helper being installed.
I had to implement this excellent change offered by @/df to get seamless WiFi connection.
https://hummy.tv/forum/threads/wifi...onally-incorrect-on-startup.9657/#post-139964
By reliable I mean it works most of the time for wireless ftp, smb, telnet, ssh, normal mode & maintenance mode.
As for your drive issue ..
I would say yes. (Is there anything of note on the SMART diags?)
As the drive may be iffy, I'd do this to try to reclaim the drive
  • remove drive - plug into a PC running Linux on a SATA3 interface (it'll be faster
:whistling:.....
Alternatively replace the drive with a brand new one (or a known/tested 100% working drive).
 
Last edited:
Back
Top