Erm, dead, or not quite?

[ptruman]

My TiVo has been occasionally found "hung up", which I suspect is due to a dodgy bit of disk .

Today however, it was stuck on "Powering up"....I rebooted it, and Cachecard kicked in, writing/verifying/caching, then it GSODs ('serious error - don't restart etc') but it reboots itself about 5 seconds later, and has been looping like that for the last coupla hours.

Is it dead? Or actually trying to recover?

[Pete77]

Is it dead? Or actually trying to recover?

Sounds like dead to me.

I think you may want to give the Emergency Kickstart procedure a go but beyond that its probably new hard drive time.

[blindlemon]

The kickstart will only result in another GSOD which the OP is already experiencing.

My money would be on serious MFS corruption - which IME is always fatal. It might be a swapfile issue, but with a 120gb drive I would consider that unlikely unless there is no swap at all.

For some reason, taking a backup with mfstools and then restoring that to a new drive will sometimes work in this situation (I expect because the process involves some manipulation of the MFS) but it's not very likely I'm afraid. You might get a minimal backup (ie. Season Passes and settings only) that works on a new drive, but again the chances are slim IME.

[ptruman]

Righto, I have a 250GB drive, MFStools 2.0 (LBA48) and a 2.5.5 image handy

Presume I want to boot the MFStools CD with

volnodma hdX=bswap

then run mfsrestore -i 255.img /dev/hdX

that should theoretically work, yes?
(possibly enabling the drive to boot first?)

I did the above before but didn't boot the bswap option and couldn't then read hdX4 or hdX7 to run the cachecard installers....

[ptruman]

Edit - rebooted with bswap on and mfsrestore works and mfsinfo picks up the drive (bizarrely on hdX10 and hdX11)

Now to get the bloody cachecard drivers running....

[Pete77]

Try www.steveconrad.co.uk/tivo/upgrade3.html for more information on formatting Tivo drives.

[ptruman]

Argh.
I now have a drive done, with 248 hours available (not tested it in the TiVo yet - but mfsinfo talks to it)

However, if I put the cachecard installer disk in and run nic_install cachecard it merrily says it can't see the TiVo hard drive....

[mikerr]

Don't forget to run copykern (option 1) as the drive is over 127Gb....

[ptruman]

I think the image I restored had the LBA48 kernel already :)

Now I have different problems - it's working, I can telnet/FTP in, I have guide data, guided setup was done, BUT loadlogos, endpad and other things are failing - any ideas?

LOADLOGOS :
TiVo: {/var/hack/scripts} % ./loadlogos.tcl ../logos/
Initialising logo list
Initialising channel list
no such object: {CONFLICT err=errTmActiveLockConflict}
while executing
"dbobj $station fsid"
("foreach" body line 10)
invoked from within
"foreach channel $channels2 {
regexp {([0-9]*)/(.*)} $channel junk id subid
set channelobj [db $db openidconstruction $id $subid]
..."
("uplevel" body line 2)
invoked from within
"uplevel $body"
invoked from within
"transaction {uplevel $body}"
(procedure "RetryTransaction" line 5)
invoked from within
"RetryTransaction {
foreach channel $channels2 {
regexp {([0-9]*)/(.*)} $channel junk id subid
set channelobj [db $db openidconst..."
(procedure "init_channelindex" line 34)
invoked from within
"init_channelindex"
(file "./loadlogos.tcl" line 583)
TiVo: {/var/hack/scripts} %

I then reran it and got something about
retrying after errTmBackgroundHoldoff ...
?? But it seems to have worked...

ENDPAD
Thursday 16:10:37 :
Thursday 16:10:37 : endpad.tcl 1.5.4 log file, starting up
Thursday 16:10:37 : Start padding set to 60 seconds
Thursday 16:10:37 : End padding set to 300 seconds
Thursday 16:10:37 : Negative start padding set to 0 seconds
Thursday 16:10:37 : Force Negative startpadding is 0
Thursday 16:10:37 : Suggestions will not be padded if another suggestion immediately follows them
Thursday 16:10:37 : Detected TiVo software version 2.5.5-01-1-023
Thursday 16:10:37 : Suggestions will be recorded in High
Thursday 16:10:37 : Timezone 00:00
Thursday 16:10:37 : Sorted configuration options (last match used):-
Thursday 16:10:37 : startpad 1 endpad 5 negpad 0 forceneg 0 maxneg -1 sugeq sugqual 75
Thursday 16:10:37 : Woken up
Thursday 16:10:37 : Error detected
Thursday 16:10:37 : can't scan path (errNmNameNotFound)

while executing
"mfs scan "/Recording/Active" -start "3:" -count 1"
("uplevel" body line 2)
invoked from within
"uplevel $body"
invoked from within
"transaction {uplevel $body}"
(procedure "RetryTransaction" line 4)
invoked from within
"RetryTransaction {
set currec [mfs scan "/Recording/Active" -start "3:" -count 1]
set nextrec [mfs scan "/Recording/Active" -start..."
Thursday 16:10:37 : Attempting to recover
Thursday 16:11:07 : Woken up
Thursday 16:11:07 : Error detected
Thursday 16:11:07 : can't scan path (errNmNameNotFound)

while executing
"mfs scan "/Recording/Active" -start "3:" -count 1"
("uplevel" body line 2)
invoked from within
"uplevel $body"
invoked from within
"transaction {uplevel $body}"
(procedure "RetryTransaction" line 4)
invoked from within
"RetryTransaction {
set currec [mfs scan "/Recording/Active" -start "3:" -count 1]
set nextrec [mfs scan "/Recording/Active" -start..."
Thursday 16:11:07 : Attempting to recover
Thursday 16:11:07 : Too many errors, stopping



Clues anyone?

[blindlemon]

volnodma hdX=bswap
If I put the cachecard installer disk in and run nic_install cachecard it merrily says it can't see the TiVo hard drive....Leave out the bswap option - you don't need it.

Just boot from the Version 1.0 LBA48 CD (http://www.tivoheaven.com/dowload/ptv-mfstools2-large-disk.iso) and run your restore - eg.

mfsrestore -s 300 -xzpi /mnt/dos/tivo.bak /dev/hdc

then run copykern, choosing option 1 to copy the LBA48 kernel and initialise the swap.

Then reboot from the SiliconDust NIC install CD (http://download.silicondust.com/tivo/nic_cd_20050218.iso) and type

nic_install/nic_install cachecard

Set your IP address etc. and shut down.

Job done :)

[blindlemon]

Now I have different problems - it's working, I can telnet/FTP in, I have guide data, guided setup was done, BUT loadlogos, endpad and other things are failing - any ideas? [...] Clues anyone?Guide data not fully indexed.

Edit: BTW, did you set a bigger swap? The default (64mb) is not large enough for a 250gb drive.

[ptruman]

I didn't specifically do anything for swap, but it appears ok currently. Can it be changed post build or not?

[ptruman]

I lie - I ran -s 127
Log shows "Adding Swap: 130044k swap-space (priority -1) "

[ColinYounger]

Seconding BlindLemon's indexing not done - errTmActiveLockConflict means that something else is using the database in such a way that you can't get access to make changes.

errTmBackgroundHoldoff means that your process will wait a bit and try again as something is making changes, but thinks it will be finished quickly. Doesn't mean it actually will be, but is trying to be friendly about it. :)

[TCM2007]

I think the image I restored had the LBA48 kernel already :)

Now I have different problems - it's working, I can telnet/FTP in, I have guide data, guided setup was done, BUT loadlogos, endpad and other things are failing - any ideas?
Clues anyone?

Let the poor dear calm down a bit. Most/all of those errors are just database busy errors. I expect its indexing the initial guide data download which can take many hours.

[blindlemon]

I lie - I ran -s 127
Log shows "Adding Swap: 130044k swap-space (priority -1) "OK, that's big enough for a 250gb drive - just - and doesn't require copykern to initialise the header.

However, you should ensure that you are running the LBA48 kernel - after a reboot, check the kernel log. There will be a line similar to the following that tells you the kernel version:-Jan 1 00:00:16 (none) kernel: Linux version 2.1.24-TiVo-2.5 (millert@xerxes.courtesan.com) (gcc version 2.8.1) #14 Wed Oct 8 12:06:25 MDT 2003 If it is as above then you do have the LBA48 kernel.

If you don't have that line but one that mentions 'build@buildmaster' instead then just pull the drive, boot from the LBA48 CD (tivoheaven.com/download/ptv-mfstools2-large-disk.iso) and run copykern choosing kernel option 1 to install the LBA48 kernel.

[ptruman]

Hmm, maybe I don't then, I see :

Jan 1 00:03:14 (none) kernel: Linux version 2.1.24-TiVo-2.5 (build@buildmaster17) (gcc version 2.8.1) #2 Thu Feb 7 17:49:48 PST 2002

[blindlemon]

That's the standard kernel. You need to run copykern ASAP.

[ptruman]

Stooopid question(s) - is there anyway to do it inline? (I know some Linuxesque implementations can do this, but unsure of this one)
and...what happens if it's not done?

[ptruman]

Right, now I'm annoyed :)

I've pulled the drive and booted with my LBA48 ISO AND the one linked above (in case they differed).

If I run copykern on that, and select Option 1 (which is DirecTivo 2.5) then I either get told "Bad boot block 0x0" (or something similar) or, if I boot and force vmlnodma and bswap, I get told "no space left on device"

Either way, if I boot the TiVo back up it works, but shows the normal (non xerxes) kernel.

Help?

[Pete77]

If I run copykern on that, and select Option 1 (which is DirecTivo 2.5) then I either get told "Bad boot block 0x0" (or something similar) or, if I boot and force vmlnodma and bswap, I get told "no space left on device"

Either way, if I boot the TiVo back up it works, but shows the normal (non xerxes) kernel.

Help?

www.steveconrad.co.uk/tivo/upgrade3.html confirms Option 1 is correct for a UK Tivo. I don't know if running for a period on a large drive without upgrading the kernel has already caused things to reach a state where it is too late to remedy them just by running copykern.

Hopefully blindlemon, mikerr or another learned member expert in upgrading drives will be along soon to advise you.

[blindlemon]

Something is definitely screwed. I would suggest re-imaging the drive from scratch but leave out the bswap option. Then run copykern immediately after the mfsrestore, and finally install the cachecard drivers etc..

BTW, did you test the drive for errors (surface scan) prior to configuring it?

[ptruman]

is bswap not important then?

[blindlemon]

You don't need to mount the drive in byteswapped mode when using MFSTools as it will do any byteswapping necessary.

Did you test the drive?

[ptruman]

No, although it is a new (bagged, sealed, unopened) Maxtor. I know it may have problems, but I didn't check it.

I did redo it all yesterday, after pulling the drive, unlocking (it had locked), restore, copykern, and install nic_config cachecard

It's now back where it appeared to be anyway, although now it has the millert@xerxes kernel showing in the logs.

It's been on overnight, guide data is loaded - but still showing odd errors?

On tivoweb "whats on" :

INTERNAL SERVER ERROR
--cut here--
action_whatson '' 'set "chnsel" "2";set "wodatetz" "1198022400";set "wotimetz" "8";set "starting" "1";set "submit" "Search";'
can't read "recstartdate": no such variable
while executing
"expr ($recstartdate * 86400) + $recstarttime"
(procedure "slotfree" line 34)
invoked from within


Also, if I search for a program, like Newsround, it finds it, but I'm getting 404 on things like http://192.168.1.200/series/3470

Whats missing?

[ColinYounger]

The image you had handy - was it a 'clean' image from WonderLander\Blindlemon\etc? Or a backup of a drive you had already?

If the image you have has MFS corruption, then obviously any restore won't help as you'll just start back with a corrupt MFS.

[blindlemon]

No, although it is a new (bagged, sealed, unopened) Maxtor. I know it may have problems, but I didn't check it.:eek: I wouldn't use any drive in a TiVo without giving it a full surface scan first! Even the best drives can suffer from shipping damage, especially when you consider how badly some of the online retailers pack them... :down:

It's been on overnight, guide data is loaded - but still showing odd errors?Could just be indexing hasn't finished.

If it doesn't clear up then try "Clear Guide Data and ToDo List" (NB. this will delete your SPs!) to clean out the old guide data and (hopefully) any associated corruption.

[ptruman]

Heh - the drive was bought a year ago when my other drive started exhibiting issues - it was well packed and has been stored on a shelf for a year, so I'm hoping it's ok - worst case, I get another one :)

[ptruman]

The image you had handy - was it a 'clean' image from WonderLander\Blindlemon\etc? Or a backup of a drive you had already?

If the image you have has MFS corruption, then obviously any restore won't help as you'll just start back with a corrupt MFS.


I bought an upgrade drive a while back (along with a cachecard) which may well have been from blindlemon or ashmeadboy (on eBay, IIRC) - which came with an LBA CD, a copy of 255.img and a CD or cachecard drivers/installs. I downloaded the latest mfstools2 iso I could find and used it to restore (remounting the cd with 255.img) and copykern - and then booted with cachecard CD to get the drivers installed.

When I went to kick off guided setup, it thought it was in SW16, if that helps :)

I have run guided setup, AND reset thumbs/suggestions etc - but I think I've had some issue running tw_setup (if that points to anything?) which may be why it's playing up?

[blindlemon]

That should be a 'clean' image as it's the one I've been using for years :)

tw_setup only installs TiVoWeb etc. and has nothing to do with guide data.

What exactly is the problem at the moment?

[ptruman]

That should be a 'clean' image as it's the one I've been using for years :)

tw_setup only installs TiVoWeb etc. and has nothing to do with guide data.

What exactly is the problem at the moment?

The above errors within TivoWeb, endpad is also giving errors.
When I'm at home I'll collate them and post them up.

It may be that I am indeed trying this too near to the rebuild/setup, and it needs time to settle...

[ptruman]

Hmm, TiVo is working and recording stuff - even the tivoguy.tcl file has bought him onto my main menu, however tivoweb is still broken.

Searching for "Transformers" gives me results, but the links fail...ie.
http://192.168.1.200/series/15839 = 404 page not found

Logos ARE going on ok - endpad still fails though - as follows from endpad.log

Wednesday 08:05:29 :
Wednesday 08:05:29 : endpad.tcl 1.5.4 log file, starting up
Wednesday 08:05:29 : Start padding set to 60 seconds
Wednesday 08:05:29 : End padding set to 300 seconds
Wednesday 08:05:29 : Negative start padding set to 0 seconds
Wednesday 08:05:29 : Force Negative startpadding is 0
Wednesday 08:05:29 : Suggestions will not be padded if another suggestion immediately follows them
Wednesday 08:05:29 : Detected TiVo software version 2.5.5-01-1-023
Wednesday 08:05:29 : Suggestions will be recorded in High
Wednesday 08:05:29 : Timezone 00:00
Wednesday 08:05:29 : Sorted configuration options (last match used):-
Wednesday 08:05:29 : startpad 1 endpad 5 negpad 0 forceneg 0 maxneg -1 sugeq sugqual 75
Wednesday 08:05:42 : Woken up
Wednesday 08:05:42 : Error detected
Wednesday 08:05:42 : can't scan path (errNmNameNotFound)

while executing
"mfs scan "/Recording/Active" -start "3:" -count 1"
("uplevel" body line 2)
invoked from within
"uplevel $body"
invoked from within
"transaction {uplevel $body}"
(procedure "RetryTransaction" line 4)
invoked from within
"RetryTransaction {
set currec [mfs scan "/Recording/Active" -start "3:" -count 1]
set nextrec [mfs scan "/Recording/Active" -start..."
Wednesday 08:05:42 : Attempting to recover
Wednesday 08:06:12 : Woken up
Wednesday 08:06:12 : Error detected
Wednesday 08:06:12 : can't scan path (errNmNameNotFound)

while executing
"mfs scan "/Recording/Active" -start "3:" -count 1"
("uplevel" body line 2)
invoked from within
"uplevel $body"
invoked from within
"transaction {uplevel $body}"
(procedure "RetryTransaction" line 4)
invoked from within
"RetryTransaction {
set currec [mfs scan "/Recording/Active" -start "3:" -count 1]
set nextrec [mfs scan "/Recording/Active" -start..."
Wednesday 08:06:12 : Attempting to recover
Wednesday 08:06:12 : Too many errors, stopping

[ptruman]

Curiouser...after loading aerialplugs logos and rebooting (twice) endpad is now not reporting errors - but /series/ links aren't working, and "Whats On" is also failing :

INTERNAL SERVER ERROR
--cut here--
action_whatson '' 'set "chnsel" "2";set "wodatetz" "1198195200";set "wotimetz" "12";set "starting" "1";set "submit" "Search";'
can't read "images": no such variable
while executing
"lindex $images [lindex $imagenumt 0]"
("foreach" body line 185)
invoked from within
"foreach showingfsid [lrange $showings $x [expr $x+19]] {
regexp {([0-9]*)/(.*)} $showingfsid junk showingobjid subobjid
set showing [db $db op..."