Doing an update from 10.3 to 11.1 the update hung about half way through. arrrrrrrrrgh!
Restarting the update produces:
error: error(-30987) setting “” records from Requireversion index
error: db4 error(-30987) from dbcursor->c_get: DB_PAGE_NOTFOUND: Requested page not found
I have tried “rpm --rebuilddb” but this does not help.
I booted from a “live” ubuntu cd to look at my hard disk. Using the ubuntu file browser, it looks like the root dir and lost and found are unreadable. My home dir and most everything else looks intact. Running fsck from the suse rescue disk does not help.
Any thoughts about how to fix this without a clean install will be greatly appreciated.
I am not sure about the fsck. I just booted the suse 11.1 disk and using the system repair tool it is telling me two things.
That the /dev/sda2 and /dev/sda3 are corrupt, but that it is unable to repair them because they are in use. I am trying to find a way to run fsck from the command line without anything else running.
I should be a little more clear. I am not sure about the fsck as it is being run by the gui and I can’t see the output. Only that the gui is now telling me that the partitions are in use and thus it is unable to fix them.
Instead of booting to the repair system try booting to the rescue system. You get a console session, no GUI. You can then login as root (no password) and run fsck manually.
I think I need to get into something like single user mode but from a CD or DVD so that I can run fsck from the command line.
From the “expert tools” menu choosing “repair file system” it reports that both /dev/sda2 and /dev/sda3 are valid file systems.
I am not sure which is telling me the truth SuSE or Ubuntu. I am suspect of the ubuntu as it is the file browser that is telling me that it cannot read the root dir.
From the Custom tools during the:
Checking package data
–Mounting all partitions
and during
searching for package database
I get a message saying:
“Changing enviorment to target system was not successful” however it does not say what system.
Yast trys to fix things but reports failing to install an rpm which I think is the kernel.
It then suggests that I need a new bootloader and offers to repair the problem that, but it fails with no explination.
Did you try to roll over the repositories? Did you attempt to upgrade from the DVD?
In any event, moving from 10.3->11.x is not a straight forward procedure. 11.0 switched to using lzma compression for rpm packages. All versions of openSUSE prior to 11.0 use bz2 and do not support lzma for rpm packages. If you want to upgrade from 10.3 you need to make sure that the upgrade goes in a specific order:
RPM (the program) is upgraded to an lzma-enabled version (11.0 or 11.1)
The remainder of the package management stack is upgraded
Then other packages get upgraded
If an upgrade starts with other things, and then attempts to unpack some rpms without the lzma-enabled package management stack…the whole process hangs since 10.3 ships with the wrong set of screwdrivers for 11.1’s screws.
Why your filesystem is reporting funny business I don’t know. But if you have a separate /home partition then I’d do the clean install. If you do not have a separate /home partition then I’d make a back-up of your /home directory and do a clean install.
Thank you for the info. At work I just updated a box from 10.0 to 11.1 using the same DVD with no problems, so I did not anticipate any here. At home things went badly. I believe /dev/sda3 is the home partition and /dev/sda2 is everything else.
Other then sftp’ing the 23GB of /dev/sda3 off to another box, is there a way to do a clean install and not touch the home partition?
If you already have a separate /home partition then you do not need to make a back-up copy of it (because you already have a back-up copy…right?).
What is important is making sure your partitioning is done correctly. During your clean install you need to opt for manual partitioning and then make sure…
sda2: Gets formatted, uses the file system of your choice, and has a mount location of “/” (the default, ‘ext3’, is fine as your file system)
sda3: Does NOT get formatted, thus retains its current file system, and has a mount location of “/home”
[Since you are formatting a partition, triple check that you are formatting the right partition. I wrote what to do with sda2 and sda3 assuming your recollection was right, make sure it is before doing anything]
So long as you create a user with the same username and UID then everything should roll over fine: files, settings and permissions. So if you were using GNOME and the username ng0g, not only will your files be there for you but your GNOME settings as well. If it is a single-user machine the UID issue should sort itself out since your UID in both installs should be “1000”.
By installing this way your home partition will be left intact and the system partition will be upgraded to the newest version of openSUSE. Having a separate home partition is how I set-up my computers. I find it to be the most flexible and trouble-free arrangement, since it allows you to keep your settings like an upgrade but also have the reduced risk of complications a fresh install provides.
ng0g adjusted his/her AFDB on Monday 25 May 2009 17:06 to write:
>
> I am wondering about “rpm --rebuilddb”. If I ran it from the “rescue”
> root prompt, am I really doing anything to my real file system?
>
>
If you are at the repair console it is possible to find the mount point of
your root on the bad machine cd into that dir, could/should be under /mnt on
the repair system.
If you can actually access the root fs with no errors you can use chroot to
change over to the machines root where you will be able to use the system
commands, you might have to use absolute paths here.
Now you could try running rebuilddb and others but I will stress here that
you could seriously and completely
b0rk your system,
before you do try this you could ( from the rescue system ) make sure that
your /home partition is not mounted so if all goes t^t`s up then it should
not touch that.
As stated this would be a last resort option, I have done it loads of times
and only lost one machine but that was down to hardware.
As always do a bit of reading on chroot and such first.
If the computer eats the cat or your hair falls out then I deny everything
as I am not really here.
even easier is to read this forum before beginning, rather than
after catastrophe…
there must be 1000 posts here saying to NOT attempt an update/upgrade
from one openSUSE version to the next, much less SKIP over one (like
11.0 in this case)…
most experienced users here recommend having /home and other data on
separate partitions (AND backed up) then format and install the new
system to root…and BE CAREFUL to not allow the new install to format
/home etc…
Taking into consideration everything everyone has contributed, it still does not account for the fact that I updated a 10.0 box straight to 11.1 with no problems the day before updating the 10.3 box.
> Taking into consideration everything everyone has contributed, it still
> does not account for the fact that I updated a 10.0 box straight to 11.1
> with no problems the day before updating the 10.3 box.
even a blind hog can find an acorn sometimes…(not that you are, but
the install script is)