Freezing on restore

[birdman]

... when it hits zero, I get the image manager screen, but now with an additional entry below the image.zip files called imagerestore (no extension)

That sounds as though the partition on which you have your imagebackups directory doesn't have sufficient free space to unzip an image, but you say there is >900MB.

Something is going wrong with the unzip, though. Do you have debug logs running (if not, enable them and try again) and what do they show?

 

[judge]

That sounds as though the partition on which you have your imagebackups directory doesn't have sufficient free space to unzip an image, but you say there is >900MB.

Something is going wrong with the unzip, though. Do you have debug logs running (if not, enable them and try again) and what do they show?

Also sounds like an issue we had in the past with some boxes not being added to image manager. Would reach the timeout with extracted image but not flash it.

 

[birdman]

Also sounds like an issue we had in the past with some boxes not being added to image manager. Would reach the timeout with extracted image but not flash it.

The supportedMachines list only has effect when you try to get a list of possible images to download. 'ixusszero' : 'Medialink-IXUSS-ZERO' is on it anyway.

A restore from an exisiting image file "just" unzips then runs ofgwrite on the result.
Hence any debug log might be useful...

 

[insertusername]

here you go

I attempted to restore the downloaded 026.zip from imagemanager
Got the imagerestore listed as described above
attempted restore of imagerestore folder
got back to image manager screen with imagerestore folder listed (no change)

 

[abu baniaz]

Can you issue the following command and post back results please?

cat /proc/mtd

 

[birdman]

Thanks for the log. It contains this:

<   423.555> [ImageManager] running commnd: /usr/bin/ofgwrite -rmtd2 -kmtd1 /media/hdd/imagebackups/imagerestore/medialink/ixusszero/
<   423.556> [eConsoleAppContainer] Starting /bin/sh
<   423.558> [eMainloop::processOneEvent] unhandled POLLERR/HUP/NVAL for fd 32(16)

(the second run is the same).

It looks as though ofgwrite isn't working?

 

[abu baniaz]

We have confirmed ofgwrite works by using it in OpenATV.

In OpenVix, Image Manager sends the parameters to ofgwrite. This is what needs to be adapted or maybe in another file that supplies the values. There are a couple of threads in the Developer section.

 

[birdman]

We have confirmed ofgwrite works by using it in OpenATV.

In OpenVix, Image Manager sends the parameters to ofgwrite. This is what needs to be adapted or maybe in another file that supplies the values. There are a couple of threads in the Developer section.

Do you mean that ofgwrite works when given the correct parameters, but OpenVix isn't sending those correct parameters?
If so - do you know what the correct ones are? And why they are different for this box (as that means that night be different for other boxes too).

 

[abu baniaz]

This is why I asked the OP to run the cpommand

cat /proc/mtd

The following threads in the bug section may help to get to the solution. These links are not available to non-staff members

http://www.world-of-satellite.com/s...-Combo-image-manager-times-out-when-restoring
http://www.world-of-satellite.com/showthread.php?38538-TM2T-couch-flashing-always-times-out

 

[insertusername]

Can you issue the following command and post back results please?

cat /proc/mtd

Sorry - am I doing this from a telnet window of from the box itself

 

[abu baniaz]

Telnet window.

Can you also confirm the contents of the extracted folder in the image restore folder please?

 

[insertusername]

telnet it is...
mtd0 07a00000 00020000 rootfs
mtd1 00400000 00020000 kernel
mtd2 00100000 00020000 complete

 

[abu baniaz]

Thanks

Try running Image manager, then post screenshot of the image restore folder please.

 

[abu baniaz]

Should be fixed in next image. Andy just committed the following.
https://github.com/oe-alliance/oe-alliance-core/commit/8e01ac2dcc1f668790e752abc1ff083c2acf7300

You must perform an online update/flash by USB until you are on 4.2 027 or newer.

 

[birdman]

Should be fixed in next image. Andy just committed the following.
https://github.com/oe-alliance/oe-alliance-core/commit/8e01ac2dcc1f668790e752abc1ff083c2acf7300

Given that mtd knows where the partitions are, why are they given as parameters to ofgwrite, rather than just let it use what it finds?

 

[insertusername]

and a screenshot - note that the problem with the device message is only because I removed the pen drive to get the log files

 

[abu baniaz]

Please check issue when next image is released.

 

[insertusername]

Thanks for all your work on this guys - admittedly I'm not sure how this strand of investigation fixes my main two problems.

1) box crashes following installation of the live football plugin - as it is without upgrade (6.1) or with upgrade (6.9) before reboot - both instances the spinner just spins rapidly upon reboot but never goes anywhere. This is during a fresh installation - no settings restore.

2) box crashes if I starting from scratch, do a basic install of cccam, add my cccam.cfg, setup ABM and search. I then create a settings backup, restore to factory settings and then restore from the recently made backup. in this instance the spinner just locks on reboot

Image install from the GUI was just another avenue I was going trying to a make a working box that I can give back to my FIL

 

[abu baniaz]

In scenario 2. why are you creating a settings restore and applying it?

I'll test the football plugin tomorrow

 

[insertusername]

Because I didnt want to keep on setting up the box from scratch when I had a problem with the box. I was testing the settings restore to see if it saved me time with the problems I was having, but because settings restore was crashing as well I was hitting a dead end.

The box seemed to be stable (with just cccam the settings , HD-red skin installed and epg and ABM setup), but after an hour I found it freezing and wouldnt boot up properly (fast moving spinner eventually locking)