CacheCard install gone horribly wrong

[webdeck]

I spent two hours trying to wedge a new CacheCard into my DSR6000, and sliced my finger in the process (it really doesn't fit in a DTiVo.) I did eventually get it to fit and booted the TiVo.

Everything worked perfectly for a couple of minutes. All the CacheCard tests passed, and Now Showing was noticably faster.

So, I then telnetted in and tarred up my files to /var, so that I could backup the working config. In the middle of that, the TiVo rebooted. The log files don't indicate why it rebooted either.

It gets to "Welcome. Powering up.", stays there a couple of seconds, then it reboots. It repeats this. I eventually unplug it.

So, I remove the cache card and reinstall the drives. Same problem.

I put the drive in my PC and take a look. Looks like /tvbin in the root partition (4) is corrupted, along with several directories in /var (9). The log files show some errors accessing hdb (but these partitions are on hda) a couple of hours before I installed the CacheCard, but there weren't any errors at all during the boot with the CacheCard installed. The log file just ends abruptly.

The alternate root (7) appears to be okay. So, I edit the boot params to point to the alternate root and kernel, and set handcraft=true so that MFS doesn't start. I figure this way, I can get it to boot to bash and I can try running e2fsck. I couldn't find any other way to run e2fsck short of building my own linux kernel with the TiVo byteswapping mods.

Only problem is, when I put the drive back in after making those changes, it continues to reboot after "Welcome. Powering up."

HELP! Is my TiVo totally hosed??? Can anyone help???

Thanks,
-Mike

[captain_video]

Check all of your cable connections to make sure you didn't dislodge one of them during the cachecard installation. The ribbon cable between the power supply and mainboard is notorious for causing problems of this sort.

[webdeck]

All cables double and triple checked - still rebooting.

-Mike

[GREEK]

take a small philips to the orange pot screw on the supply and tweak it a 1/8 of a turn, usually up works best, may try down after up doesnt work. Usually helps this problem out. Otherwise you may have to yank the cache card and reimage to make sure something else isnt going on, or use a spare drive if available.....

[webdeck]

I found the KNOPPIX boot CD, which has e2fsck. So, I booted with that and ran e2fsck, and as it was repairing the filesystem, it reported a lot of hard failures reading some sectors on the drive, so several files in /tvbin and /var were not recoverable.

So my plan is now to clone the drive onto another one (which arrives on Friday), and then copy over the missing files in /tvbin from another TiVo, then try and boot it again (the TiVo should rebuild /var on its own.)

Any recommendations for the best way to clone a drive with bad sectors?

Thanks,
-Mike

[webdeck]

Okay, so I got a new 160GB drive, and copied the old drive to it. The copy ran into some errors. The new drive's root partition appears to be okay, except for the missing /tvbin directory. The new drive's /var partition was fairly hosed, so I just blew it away and gave it a blank one.

I don't have another DSR6000 running 3.1.0b-01-1, but I do have /tvbin in the alternate root partition, which is from 3.1.0-01-1-001, so I copied that over, hoping it would be close enough for now, until I could extract the correct cpio from MFS.

So, it definitely booted further this time. I get to the dreaded green screen, it stays there for a couple of seconds, and then it reboots. I've attached the relevant log files.

Now, I've posted in a separate thread to see if someone can get me the correct /tvbin contents, in the hope that the version difference is why it is failing.

I really don't want to lose the hundreds of unwatched recordings on this TiVo. If anyone knows how I can get it up and running again with the shows intact, please let me know!

Thanks,
-Mike

[AVD]

Cachecard caches the whole MFS database. So it reads parts of the disk that were never accessed. So it finds errors that already existed on hte harddrive that were never apparent before.

It sounds like your drive is dying, change it.

(also make sure the drive is mounted read-only before turning off power)

[webdeck]

So, both Maxtor 160GB drives had bad sectors. I purchased two new 160GB drives (a Seagate and a Hitachi) and used dd_rescue to copy over as much as I could. There were 25 bad sectors in one of the MFS Application partitions, so the TiVo then went into the GSOD for several hours. It came out of the GSOD and seems to work and have all its recordings. I made a backup and then reinstalled the cachecard and all appears to be okay now. Both Maxtor drives are still under warranty, so they're going back for replacement.

The key to success with a hard drive with bad sectors is to use dd_rescue - trying to use dd will miss too much data, and will take much longer.

-Mike