[SOLVED] Online updating SuSE: kernel rpm could not be installed

JZL240I-U · 04-09-2010, 02:30 AM

There was a kernel update included in my last three attempts of online updating SuSE 11.2 -- all of which which ended with an error, something like kernel rpm could not be installed (one can't copy that out of the YAST-window

).

The new kernel-relevant directories are there under /boot, the original 2.6.31.5 is gone. This produced lots of errors during the next boot until I manually re-assigned links in /lib/modules pointing to the new kernels. "uname -a" shows my old kernel (see also this thread: http://www.linuxquestions.org/questi...46#post3927146).

I ran "rpm --rebuilddb" and re-tried the installation, all to no avail.

In the message box during install came about 1,5 MB of errors, a brief extract is shown below:

Code:

Subprocess failed. Error: RPM  fehlgeschlagen (<- failed): Free diskspace below /boot: 3782912  blocks
error: db4 error(-30987) from dbcursor->c_get: DB_PAGE_NOTFOUND:  Requested page not found
error: error(-30987) getting "ddecac6d" records from Provideversion  index
error: db4 error(-30987) from dbcursor->c_get: DB_PAGE_NOTFOUND:  Requested page not found
error: error(-30987) getting "f5011f67" records from Provideversion  index
error: db4 error(-30987) from dbcursor->c_get: DB_PAGE_NOTFOUND:  Requested page not found
....
error: db4 error(-30977) from dbcursor->c_get: DB_RUNRECOVERY: Fatal  error, run database recovery
error: error(-30977) getting "" records from Filedigests index
rpmdb: PANIC: fatal region error detected; run recovery
error: db4 error(-30977) from dbcursor->c_get: DB_RUNRECOVERY: Fatal  error, run database recovery
error: error(-30977) getting "+��fF� V � �   " records from Filedigests  index
rpmdb: PANIC: fatal region error detected; run recovery
....
rpmdb: PANIC: fatal region error detected; run recovery
error: db4 error(-30977) from db->sync: DB_RUNRECOVERY: Fatal error,  run database recovery
rpmdb: PANIC: fatal region error detected; run recovery
error: db4 error(-30977) from db->sync: DB_RUNRECOVERY: Fatal error,  run database recovery
....

Any ideas how to proceed? Can one delete the rpm database and rebuild it from the online repositories?

ButterflyMelissa · 04-10-2010, 03:10 PM

Hi,

Ok, just of the top o' my head, here...but it is possible to start the old kernel, assuming that you cannot launch it - if I hit the ball wrong, sorry...

Go su and edit /boot/grub/grub.conf (I assume that SUSE has a Grub, I use Fedora) to allow some timeout (it says timeout=0, possibly, edit that to the amount of seconds you need to select and launch the required kernel) and commend the "hidemenu" out. Restart and select the kernel previous to the "update" (using that word loosely - absolutely no malicious pleasure intended here, believe me) - that should get the girl up and running. Now, to remove the kernel in question. Have a look at this page
http://www.cyberciti.biz/faq/debian-...ernel-command/
I hope this helps.

Wellness to you!

Thor

JZL240I-U · 04-12-2010, 01:58 AM

Quote:

Originally Posted by linusr@flanders

...but it is possible to start the old kernel, assuming that you cannot launch it...

I can start it, no sweat, the system is running okay. If you read the other thread (see link in my first post), you will see that it is rather the problem of updating my system.

Quote:

Originally Posted by linusr@flanders

...Restart and select the kernel previous to the "update" (using that word loosely - absolutely no malicious pleasure intended here, believe me) - that should get the girl up and running.

Not necessary, but thanks anyways.

Quote:

Originally Posted by linusr@flanders

...Now, to remove the kernel in question. Have a look at this page
http://www.cyberciti.biz/faq/debian-...ernel-command/
I hope this helps...

I'll try this and cross my fingers, but I would like to know, why I get an error when I try to update my system.

ButterflyMelissa · 04-12-2010, 03:22 AM

Hi,

Hmm, I AM a new-be (only three-odd years, but Linux keeps surprising, so, one is always a newbe somewhere) so sorry abut wasting time...did find this:

Quote:

http://www.linuxquestions.org/questi...errors-337269/

I saw the phrase "

Quote:

Fatal error, run database recovery

", so that could help, furthermore, I noticed this:

Quote:

"Requested page not found error: error(-30987) getting "ddecac6d" records from Provideversion index"

And I had this too, on my F8 going to F12. It could (*could*) mean that a server is either (temp) down or a repo is off-line/obsolete. You may have to tinker with your repos (Suse must have a folder like yum.repos.d where all repos are located) - each repo (a location where software & updates are stored on the net) has its own file, I'd find the one with the text "Provideversion" in it, and set it

Quote:

enabled=0

- at least for a little wile...

I got that answer at this page:

Quote:

http://www.question-defense.com/2009...-and-try-again

But...as I mentioned before, I "sniff" Fedora...

Wellness 2 U!

Thor

JZL240I-U · 04-12-2010, 03:49 AM

Quote:

Originally Posted by linusr@flanders

...Hmm, I AM a new-be (only three-odd years, but Linux keeps surprising, so, one is always a newbe somewhere) so sorry abut wasting time...

Aren't we always?

I did already try "rpm --rebuild" but probably not as root

. Will try soonest (tomorrow) again (first looking for /var/lib/rpm.old. If that is not there the problem seems pinpointed).

I doubt that it is a repository error. I repeatedly tried to install the rpm without downloading it anew and also with many weeks in between to tries of downloading and installing it.

JZL240I-U · 04-14-2010, 01:59 AM

Darn. I did it as root. But I can't find that rpm-database. There is a /var/lib/rpm/* directory with 33 items in it, but none look like it and there is no .old file either. Grrr

.

Next I did a "rpm -qa | grep kernel" and got -- tataaa -- the newest and latest kernel which is most regrettably

not installed. And I had not idea at all how to manually try to install it again. How does one install a packet which is already in the rpm database?

JZL240I-U · 04-30-2010, 07:03 AM

Umm. Would a

Code:

rpm -i -f <new.kernel.rpm>

work while the old kernel is still in command of the OS?

JZL240I-U · 06-03-2010, 04:42 AM

To answer my own question: yes, it would work (I just remembered doing this years and years ago). But it would overwrite (or possibly install the new along) the old kernel with no further notice.

To avoid this, a true installation is necessary, since an installation (or rather YAST) adds a stanza for the new kernel to the entries in /boot/grub/menu.lst it leaves the old kernel and its libraries intact (just creating symbolic links) and so can be undone if necessary.

The reason behind all this trouble:

I had two /boot directories (and even partitions). During system start up I used the old /boot of SuSE 11.0 but in the /etc/fstab I had an entry for the other and new /boot of SuSE 11.2 ... accordingly I updated that (the new one) while starting from the other (the old one) and thus never got the new kernel where it belongs.

I don't quite understand why the old /boot got the libraries and modules but not the kernel. Anyway, after I corrected the /boot entry in /etc/fstab everything is as it should be

. If anybody is interested here are more threads connected to this topic:

http://www.linuxquestions.org/questi...formed-806910/
http://www.linuxquestions.org/questi...t-from-796986/

Thanks for your support and help Thor

.