I think HDD has died

[wbeasley]

Well, has been freezing daily lately with pixels on screen, used to happen every few months.
Last night It gave an error message on reboot about possible corrupt HDD when starting to populate cache...so figured ahhh lets reimage

Reimged over old possible corrupt hard drive and got different error message when trying to detect cachecard...possible similar

Brought new 320GB HD today , imaged (aug TWP latest image), now on bootup is just says "please wait for TIVO....message"
(Edit This was because drive was cable select...)

I did notice when I copied image to HDD that some errors came up, possible corrupt image?? Next step will be to reimage after Dinner but any thoughts would be appreciated

cheers

[wbeasley]

Is it normal to get some crc errors when coping image at the start? I guess not but ijhave a brand new 320GB Hitachi HD, seems a coincidence both old drive and new drive get the same errors?

when I boot tivo the cachecard drivers do not load, is there something different in this image? May try the earlier image instead

[wbeasley]

Well, decided to keep persevering and the cachecard driver does not load however network is working and guide data has downloaded, i'm going to let tivo do it's thing and see if it's stable or not.

I remember the same driver issue on an old image and when i reimaged last year the cachecard worked for the first time ever

[Darren King]

Download the CD bootable version of Seatools from www.seagate.com and run a full scan on the suspect hard disk. This program will work even if the hard disk is not a Seagate brand.

All the symptoms you have mentioned are caused by a dying hard disk. The average lifespan of a TiVo hard disk is around 30 months or so and given the hard disk is always in use (even when you are not recording or watching anything) this is a good lifespan.

Sorry can't help with the cachecard issue.

[petestrash]

I suggest you run the seatools program Darren suggested on your new drive also. CRC errors are not normal.

I don't use the NZ image, but I recall a similiar issue in an OzTiVo version a while ago. Post a copy of your rc.sysinit file and I'll take a look.

Peter.

[NZMarke]

I had very similar symptoms to this (also with a Seagate). I used the Seagate utility to check it, and it reported no problems. So I replaced the disk again. Same issues showed up. I then replaced the disk completely with a brand new (and larger) model, and it's worked flawlessly since. As has the old "faulty" disk - which I put into an old desktop for my wife (grin) as a secondary drive - so go figure.

My theory - for what it's worth - is that for the price of a new disk, it's not worth hassling about with checking the old one, finding no errors and replacing/reimaging only to find the same old problems reoccur when it's reinstalled on the TiVo. Better off to just bite the bullet and put a new disk in. The advantage is that you can put a bigger one in!! ;-)

Cheers

Mark

[NZMarke]

And realised you have two discs - one old, one new. So have you tried just running the newer one - in case errors on one are causing the problem on both?

[wbeasley]

Thanks for the replies

The new disk has been stable since I got it working, unsure what the crc errors were when I was copying the image accross...

The cachecard driver is still not working, doesn't load, some buffer type errors (sorry going from memory) also populates message something like not being able to access the kernel. Had these same errors on the origional image i used a few years back.

Thanks petestrash, I Will post rc.sysinit when i get a chance over the weekend

[petestrash]

Chances are the previous image you used was an OzTiVo one, as I believe this is the first NZ image release.

Peter.

[wbeasley]

#!/bin/bash
# Copyright (c) 2001, 2002 TiVo Inc.
#
# /etc/rc.d/rc.sysinit - run once at boot time
#

#
# Useful functions
#

function nthword () {
n=0
for foo in $* ; do
if [ $n -eq $1 ]
then
echo $foo
fi
n=$(($n+1))
done
}

function runme () {
echo "Scanning for $1 repair scripts"
for foo in /var/persist/*.$1 ; do
if [ -f $foo ]
then
echo "Running repair program $foo"
rm -f $foo.old
mv $foo $foo.old
chmod 555 $foo.old
if /tvbin/crypto -vfs $foo.sig $foo.old /tvlib/misc/service-v3-s.pub ; then
$foo.old || echo "$foo failed."
else
echo "$foo failed signature check."
fi
rm -f $foo.old $foo.sig
fi
done
}

function fixmodem
{
typeset MAXTRIES=3

typeset MODEMTYPE=/tvlib/modem/utils/modemtype.expect

if test ! -r $MODEMTYPE ; then
echo "fixmodem: no modem id program: $MODEMTYPE"
return
fi

typeset modem_name=$(/bin/expect $MODEMTYPE)

if test -z "$modem_name" ; then
echo "fixmodem: no modem name"
return
fi

typeset PATCHER=/tvlib/modem/patches/$modem_name/ram/expect_script

if test ! -r $PATCHER ; then
echo "fixmodem: no patcher: $PATCHER"
return
fi

#
# There is a patcher.
# Let's see how many times we have tried to use it.
# The file is supposed to consist of a single line; the number
# of characters on that line is the number of tries so far.
#
typeset FLAGFILE=/var/persist/modem_patch_tries

#
# The FLAGFILE must be a regular file, and it must be readable and writable.
# Otherwise, the file is created from scratch.
#
test -f $FLAGFILE -a -r $FLAGFILE -a -w $FLAGFILE ||
{
rm -rf $FLAGFILE
echo 'X' > $FLAGFILE
}

echo "Attempting to fix modem using: $PATCHER"

/bin/expect $PATCHER

typeset result=$?

if test $result -ne 0 ; then
typeset line tries

read line < $FLAGFILE
tries=${#line}

if test $tries -lt $MAXTRIES ; then
echo "Failed to fix modem; this was try #$tries"
line="${line}X"
echo $line > $FLAGFILE

#
# Verify that the file size has actually changed
#
typeset new_line

read new_line < $FLAGFILE
if test "$new_line" = "$line" ; then
sync
sync
sync
echo "Rebooting system to attempt another modem patch"
reboot
fi
echo "Failed to increase try count (file system full ?)"
else
echo "Too many unsuccessful tries to patch modem: $tries (giving up)"
fi
fi

#
# Eliminate the FLAGFILE, so that next time we reboot, we start
# the whole patch process again.
#
rm -f $FLAGFILE
}

# Pull in arch-specific bits
source /etc/rc.d/rc.arch

# Set the path
PATH=/devbin:/bin:/sbin:/tvbin:/hack/bin
export PATH

# Read in our testing configuration, if there is one.
[ ! -f /test.conf ] || source /test.conf

# Some tcl scripts expect TIVO_ROOT to be set. It would be cleaner to
# just use the path, but that's not the way it is right now. The
# contents of TIVO_ROOT is prepended to paths, so the empty string is
# just fine.
TIVO_ROOT=
export TIVO_ROOT

if [ "$sysgen" = true ]; then
echo
echo
echo Starting shared library installation environment
echo You may Telnet in. The telnet connection
echo will run bash as the login shell.
echo

configEtherSysinit

echo Starting Telnet Listner ... >& /dev/console
/sbin/tnlited 23 /bin/bash >& /dev/console

echo Starting /proc Listener ... >& /dev/console
mount -n /proc /proc -t proc >& /dev/console
procd >& /dev/console

exit
fi

echo "Starting rc.sysinit"

# Release the old initrd
echo "Releasing /initrd and clearing ramdisk, if they exist"
umount -n /initrd
freeramdisk /dev/ram

# Start up swapping.
echo "Activating swap partitions"
swapon -a

echo "Loading i2c driver"
insmod /lib/modules/i2c.o

echo "Loading core system drivers"
loadCoreAndSetTV_STD

loadIRDriver

# This should be read from /etc/fstab... or a config file somewhere.
export varpartition=/dev/hda9
if pdisk -l /dev/hda |grep "Bootstrap 1" |grep 128.0M ; then
echo "Partition 2 is 128MB. It must be the hack partition"
export hackpartition=/dev/hda2
else
echo "Partition 2 is not 128MB. The hack partition must be 4 or 7"
echo "Checking for root partition"
if [ "$root" == "/dev/hda4" ] ; then
echo "Hack is on partition 7"
export hackpartition=/dev/hda7
else
echo "Hack is on partition 4"
export hackpartition=/dev/hda4
fi
fi

# Check for panic signal
echo "Checking for Kickstart panic signal"

checkpanic

panic=$?
export EMERGENCY_REINSTALL=0
do_mfs_assert=0
do_mfs_cleanup=0

if [ $panic -ne 0 ] ; then
case $panic in
169)
echo "Kickstart code 5 2 - emergency reinstall"
export EMERGENCY_REINSTALL=1
;;
173)
echo "Kickstart code 5 6 - software install"
export swupgrade=true
;;
174)
echo "Kickstart code 5 7 - force MFS check"
do_mfs_assert=1
;;
175)
echo "Kickstart code 5 8 - perform MFS cleanup"
do_mfs_cleanup=1
;;
*)
kickstart $panic
reboot
;;
esac
fi

echo "Loading input section drivers"
loadInput

echo "Cleanup $varpartition pass 1"
if e2fsck -p $varpartition ; then
echo "$varpartition is clean"
else
echo "Cleanup $varpartition pass 2"
if e2fsck -p $varpartition ; then
echo "$varpartition is clean after pass 2"
else
echo "Can't clean $varpartition - rebuilding"
mke2fs -c $varpartition
echo "Mounting /var to rebuild it"
mount -t ext2 -n $varpartition /var
mkdir /var/log /var/tmp /var/run /var/packages /var/dev /var/bin /var/utils /var/persist
touch /var/timestamp
ln -s /hack /var/hack
umount -n /var
fi
fi

echo "Cleanup $hackpartition pass 1"
if e2fsck -p $hackpartition ; then
echo "$hackpartition is clean"
else
echo "Cleanup $hackpartition pass 2"
if e2fsck -p $hackpartition ; then
echo "$hackpartition is clean after pass 2"
else
echo "Can't clean $hackpartition - Pull the drive and run e2fsck manually."
fi
fi

# If this is an NFS root, we may want to allow for /var to stay in root
mount -v -n /var

### DO NOT CHANGE THIS UNLESS YOU KNOW WHAT YOU ARE DOING.
mount -t ext2 $hackpartition /hack

echo "Cleaning up /var/mtab..."
cp /dev/null /var/mtab

# Set the TIVO_SVR_ADDR up.
if [ ! "$xTIVO_SVR_ADDR" = "" ]; then
echo "Changing SVR ADDR to $xTIVO_SVR_ADDR"
export TIVO_SVR_ADDR=$xTIVO_SVR_ADDR
else
echo "Using default TIVO_SVR_ADDR"
export TIVO_SVR_ADDR=192.168.50.1:80
fi