Results 1 to 11 of 11

Thread: HELP! - Stuck on the "Almost There' screen

  1. #1
    Join Date
    Jul 2003
    Location
    Raleigh, NC
    Posts
    19

    HELP! - Stuck on the "Almost There' screen

    Help! I'm stuck at the 'almost there' screen. I know there is really no point in describing what the symptom are, without, first describing what I did, so here is goes...

    I finally decide to bite the bullet and upgrade my sleepered 3.1.1c Tivo to 6.2. I modified the installSW.itcl scripts, ran it, dd the new killinintrd kernel (Or so I thought - more later...), and re-hacked (AW utiltiies, rc.sysinit.author, etc...). I did NOT flip the bootpage, because the installSw script did that for me. And then perhaps the the source of all my problems, in a weak moment I entered 'restart' instead of rebooting. Well, none of my hashes match so it deleted everything and then stopped with a kernel panic: no init found. (At the Welcome... Powering up' screen) Crap! So, after much reading, I found two references to threads PlainBill has responded to about the the absence of a killinitrd kernel will cause my symptom. .. but no solution was offered, only 'been there done that' ... more reading. So, at this point, I still don't know if my mistake was the 'restart' command or the kernels weren't dd right. No problem, I'm not too stresses at this point... keep on reading. I then came across other threads that appreared to have similar issues. Their solution was to copy all the root files from a know good config to the bad one. Well, I already have another hacked 6.2 box upstairs, so I copied all of it's files over to my broken tivo. ... and of course, redid the kernels.

    So, I have seemed to have fixed the kernel panic error, but now I'm stuck on the 'Almost there' screen. I can see from my console log, rc.sysinit is completely finished, but nothing else happens! I ran 'ps aux' thinking something was running, nothing looked obvious. Can someone please take a look? I'm really at a loss. I have attached all the console log of the installSW.itcl, the restart, and then the console of the boot where I get stuck. What is happening? Is something running, or not running... I'm not sure what comes are the rc.sysinit completion. Is there something else I did wrong? Your help is greatly appreciated.

    I still have console access, so is there something I can run? Is it adviseable to run the installSw.itcl script again? Of course, this time, the 6.2 version of the script would get run. Would this be a problem running a 6.2 installSw on top of a 6.2 load?

    Thanks!

  2. #2
    Join Date
    Feb 2002
    Posts
    6,413
    Dude, you're probably going to have to start over from scratch with a fresh image. You should have unsleepered the unit before attempting to upgrade to 6.2. You started with a monte'd setup and tried to do a normal upgrade. You're probably trying to boot into an incompatible kernel due to the way things were set up. I would advise installing a clean image to the Tivo and take it from there. If you Sleepered the unit it should have created a backup image for you. Just restore the image using the tivoscripts CD but don't do the monte or hacks phases.
    Please don't PM me or any other members looking for personal assistance. You'll do better by posting (after you've exhausted the search feature, of course) and taking advantage of the collective expertise of the membership instead of a single individual that may or may not be able to help you. Thank you and enjoy your stay at DDB!

  3. #3
    Join Date
    Jan 2002
    Posts
    5,601
    I'm not sure if it is possible to recover from this without installing a fresh image. installSw.itcl used the old boot parameters. As a result it is trying to run kmonte.o. That is SURE to cause problems.

    A quick review of the first two logs point out a couple of problems. The most important: Your screwup resulted in the following boot paramaters: root=/dev/hda4 dsscon=true console=2,115200 BASH_ENV=`mount$IFS-n$IFS/dev/hda16$
    IFS/mnt;echo$IFS/mnt/runmonte`
    THAT is incorrect. Change it to: root=/dev/hda4 dsscon=true console=2,115200 upgradesoftware=false

    While you're making the change, what does 'bootpage -pb /dev/hdc' report? (assuming you have the drive hooked up as secondary master). It may also be necessary to correct /etc/fstab.

    PlainBill
    Last edited by PlainBill; 08-16-2005 at 11:20 AM.
    There's a difference between needing help, and just being plain ole' lazy.

    "You cannot teach a man anything. You can only help him find it for himself." Galileo Galilei (1564-1642)

    HR20-700 with 2 TB, HR22-100, HR22-100, HR22-100, HR23-100 all running 0x5cd and networked.

  4. #4
    Join Date
    Jul 2003
    Location
    Raleigh, NC
    Posts
    19
    Thanks for replying ... continuing to try to salvage the mess I've currently created... is it possible to flip the bootpage, killinitrd a 3.1.1 kernel, and go back to 3.1.1c, or did the upgrade overwrite the old image?

    That aside... Plainbill, you said I used the old monte boot parms when installSw was run. Yes, that is correct. But, I did change it before I restarted. I assume there is a difference? ...just trying to learn a few things here.


    when I ran (before the restart) 'bootpage -p /dev/hda' it returned the following:
    root=/dev/hda4 dsscon=true console=2,115200 upgradesoftware=false
    I know this for sure.

    Unfortunately, I didn't initially do the checks for 'bootpage -b' and 'etc/fstab' before I did the restart, so I'm trying to recall what it did say when I finally did run the commands.... I'm doing this from work too, so I'm going on memory. But, I think at one point 'bootpage -b /dev/hda' returned a 7, and /etc/fstab returned 4, if that makes any sense.

    Anyway, (after some more reading) I realized this was wrong. I know for sure, 'bootpage -b /dev/hda' returns a 4 now (I'm using the i386 version of bootpage.. so the real one should be 3, right?). and '/etc/fstab /dev/hda' return 4.

  5. #5
    Join Date
    Jan 2002
    Posts
    5,601
    I'n not sure you're screwed, but I don't see a lot of options. Here are a few suggestions:

    Since you've got console access, compare the contents of /etc/fstab, bootpage -b and bootpage -p (the mips version of bootpage). Make sure they all point to the 3/4 partition set.

    Did the database upgrade run? If not, that could be the source of your problem. Trace through /etc/rc.d/rc.sysinit and figure out where it should have run and see if you can trigger it.

    Edit installSw.itcl to remove the 'sanity check' and reinstall 6.2. I have problems following tcl scripts, so you're going to be on your own.

    Edit installSw.itcl and install 3.1.1x (whatever version you sere running - if the slices are still present).

    Write it off to experience and reimage.

    PlainBill
    There's a difference between needing help, and just being plain ole' lazy.

    "You cannot teach a man anything. You can only help him find it for himself." Galileo Galilei (1564-1642)

    HR20-700 with 2 TB, HR22-100, HR22-100, HR22-100, HR23-100 all running 0x5cd and networked.

  6. #6
    Join Date
    Jul 2003
    Location
    Raleigh, NC
    Posts
    19
    Thanks PlainBill for the assistance. I'll give your ideas a shot tonight when I get home.

    A few questions though.. when you said to remove the 'sanity check' in the installSw file. Is that the exact verbage I should look for? Just curious... I know you said you weren't that familiar with TCL, but thought I would ask for some clarification.

    Anyway, I think first thing when I get home I'm going to re-verify all the bootpage -p, bootpage -b, and etc/fstab all report the 3/4 partition. Then, I'm going to re-copy the partition from my other Tivo to my Broken Tivo. Just to verify I used the right commands (I've done so many things wrong already):
    Code:
    dd if=/dev/hdc4 of=/dev/hda4
    where hdc is my GOOD Tivo (I assuming the partition is 4 for my little example) and hda is my broken Tivo.

    A few other things... my GOOD Tivo is is a 2-drive hacked system. Is this a problem? I'm pretty sure I made a backup 6.2 image somewhere (At least, I hope!)

    Going back to a previous question, I assume the old 3.1.1c partition is no longer valid? Just wondering, if I could flip the bootpage back (and just bootpage -p), killinitrd, reboot, and then start over with the 6.2 upgrade.

    With regards to review rc.sysint, are there any logs I can look at? /var/log/kernel/log?

    Thanks again for your help.

  7. #7
    Join Date
    Jan 2002
    Posts
    5,601
    Quote Originally Posted by bluemodena
    Thanks PlainBill for the assistance. I'll give your ideas a shot tonight when I get home.

    A few questions though.. when you said to remove the 'sanity check' in the installSw file. Is that the exact verbage I should look for? Just curious... I know you said you weren't that familiar with TCL, but thought I would ask for some clarification.
    No, it's not. If I read the script properly, by changing set emergency 0 to set emergency 1, it may do the reinstall. Frankly, I'd wait for confirmation before trying this.

    Quote Originally Posted by bluemodena
    Anyway, I think first thing when I get home I'm going to re-verify all the bootpage -p, bootpage -b, and etc/fstab all report the 3/4 partition. Then, I'm going to re-copy the partition from my other Tivo to my Broken Tivo. Just to verify I used the right commands (I've done so many things wrong already):
    Code:
    dd if=/dev/hdc4 of=/dev/hda4
    where hdc is my GOOD Tivo (I assuming the partition is 4 for my little example) and hda is my broken Tivo.
    That's right. To simplify things, check the boot parameters on the good drive and move the active partition set to the same partitions on the 'bad' drive; then update the boot parameters on the bad drive.
    Quote Originally Posted by bluemodena
    A few other things... my GOOD Tivo is is a 2-drive hacked system. Is this a problem? I'm pretty sure I made a backup 6.2 image somewhere (At least, I hope!)
    The software partitions are not changed when you add a second drive; the changes are in the mfs partitions.
    Quote Originally Posted by bluemodena
    Going back to a previous question, I assume the old 3.1.1c partition is no longer valid? Just wondering, if I could flip the bootpage back (and just bootpage -p), killinitrd, reboot, and then start over with the 6.2 upgrade.
    Quick background: A stock TiVo has two partition sets 2/3/4 and 5/6/7. (Partitions 2 and 5 are unused.) One set is active, the other set is the alternate. Bootpage -b is used to indicate the boot (kernel) partition, bootpage -P to indicate the root partition. The two partition monte installed by TiVoScripts uses both partition sets. The one active in bootpage is actually an obsolete, but hackable one. It chain loads the second partition set, whick is your hacked operating system. Nutkase has a better description here.

    If I understand what actually happened, installSw.itcl installed the new root partition over the old hacked partition set. The partition set that is still present is the obsolete one. Even if the good 3.1.1c root file system were present and you replaced the kernel with a killhdinitrd one, it would barf (a technical term) when it accessed the mfs database and saw it was configured for 6.2.
    Quote Originally Posted by bluemodena
    With regards to review rc.sysint, are there any logs I can look at? /var/log/kernel/log?

    Thanks again for your help.
    PlainBill
    There's a difference between needing help, and just being plain ole' lazy.

    "You cannot teach a man anything. You can only help him find it for himself." Galileo Galilei (1564-1642)

    HR20-700 with 2 TB, HR22-100, HR22-100, HR22-100, HR23-100 all running 0x5cd and networked.

  8. #8
    Join Date
    Jul 2003
    Location
    Raleigh, NC
    Posts
    19
    Quote Originally Posted by PlainBill
    No, it's not. If I read the script properly, by changing set emergency 0 to set emergency 1, it may do the reinstall. Frankly, I'd wait for confirmation before trying this.
    Is this the thread you were thinking of? JeffD does it slightly different... does 'if {1}' which would always be true.

    Quote Originally Posted by PlainBill
    Quick background: A stock TiVo has two partition sets 2/3/4 and 5/6/7. (Partitions 2 and 5 are unused.) One set is active, the other set is the alternate. Bootpage -b is used to indicate the boot (kernel) partition, bootpage -P to indicate the root partition. The two partition monte installed by TiVoScripts uses both partition sets. The one active in bootpage is actually an obsolete, but hackable one. It chain loads the second partition set, whick is your hacked operating system. Nutkase has a better description here.

    If I understand what actually happened, installSw.itcl installed the new root partition over the old hacked partition set. The partition set that is still present is the obsolete one. Even if the good 3.1.1c root file system were present and you replaced the kernel with a killhdinitrd one, it would barf (a technical term) when it accessed the mfs database and saw it was configured for 6.2.
    Ah... I see. It makes sense why things aren't rebooting correctly. I'm just hoping I can learn how to correct my mess, save my recordings and hopfully learn something along the way. Nobody has said anything about the MFS database WRT it being corrupted. So, I'm hoping I'll will be able to watch my shows if/when it finally boots. <fingers-crossed>


    I was doing a little more reading before I go try to recopy the root partition over again. I found a few links about NOT upgrading installSW with a Sleeper'd drive. Oops!

  9. #9
    Join Date
    Jan 2002
    Posts
    5,601
    Quote Originally Posted by bluemodena
    Is this the thread you were thinking of? JeffD does it slightly different... does 'if {1}' which would always be true.
    I have sucessfully installed 3.1.1c over 3.1.1c by removing the ! (IF I remember correctly ). However, installSw.itcl has changed in 6.2, so I'm fumbling around with this.

    Quote Originally Posted by bluemodena
    Ah... I see. It makes sense why things aren't rebooting correctly. I'm just hoping I can learn how to correct my mess, save my recordings and hopfully learn something along the way. Nobody has said anything about the MFS database WRT it being corrupted. So, I'm hoping I'll will be able to watch my shows if/when it finally boots. <fingers-crossed>
    The mfs database should not have been corrupted - but I'm sure that we can do that if we fool around enough. Incidentally, one thing you might want to do before doing anything too drastic is to boot the 'good' system, capture the serial console output, and compare it to the 'bad' system.

    Quote Originally Posted by bluemodena
    I was doing a little more reading before I go try to recopy the root partition over again. I found a few links about NOT upgrading installSW with a Sleeper'd drive. Oops!
    Yup. I could have told you that.

    PlainBill
    There's a difference between needing help, and just being plain ole' lazy.

    "You cannot teach a man anything. You can only help him find it for himself." Galileo Galilei (1564-1642)

    HR20-700 with 2 TB, HR22-100, HR22-100, HR22-100, HR23-100 all running 0x5cd and networked.

  10. #10
    Join Date
    Jul 2003
    Location
    Raleigh, NC
    Posts
    19
    I took a better look at the boot log. It looks like the system can't execute files like TvLauncher (which seem important)! This is symlink to tivoapp. I even tried running tivosh, says it can't execute.

  11. #11
    Join Date
    Jul 2003
    Location
    Raleigh, NC
    Posts
    19
    Well, I have great news. After to re-dd'd the root partition over from a good image. Everything booted fine. It took the upgrade, got past the "almost there" screen and everything came back. And all my shows are intact! Yeah!!

    For PlainBill... Thank you Thank you Thank you Thank you!!!!!

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •