TypeError: rpmdb open failed

[u-noneinc-s]

the command below gives me an ERROR

[root@localhost ~]# rpmdb --rebuilddb
rpmdb: PANIC: fatal region error detected; run recovery
error: db4 error(-30977) from dbenv->open: DB_RUNRECOVERY: Fatal error, run database recovery
error: cannot open Packages index using db3 - (-30977)

Did you remove the db before you used the rebuild command?

# rm /var/lib/rpm/__db*
# rpm --rebuilddb
# yum -y update

The repos are in /etc/yum.repos.d although I'm not sure you have a problem with them. Yesterday you said it was updating.

Note, that is 2 underscores before db* in #rm /var/lib/rpm/__db*

[hermouche]

ok i got it .

thanks

red

[meverhagen]

Yeah a confusing yum flaw
I can run yum for several times without error, but sometimes for no obvious reson it gives an error.
The six repo's I use aren't the problem. The correct ones are enabled and the lots of others aren't.

It his hard to say what is causing it.

this is the error notice:

rpmdb: unable to allocate space from the buffer cache
error: cannot open Packages index using db3 - Kan geen geheugen toewijzen (12)
error: cannot open Packages database in /var/lib/rpm
Component: Software Updater
Summary: TBc3704fb6 config.py:656:_getsysver:TypeError: rpmdb open failed


I found two ways of getting yum to run again,

the first is to delete the db (No need to rebuild with rpm --rebuild, this is nessesairy when yum fails on lib dependacies wich are present on your system, wich occasionaly happens with self compiled things.)

# rm /var/lib/rpm/__db*


the second is to reboot.

I wonder, does the reboot clean up the db when shutting down ?

One of the db files __db.000 appears to be empty, a file of zero lenght.
I think this empty file is the somehow the problem. (Yum can;t open an empty db index, cause it doens't exists ! flaw)

Yum seems to create the empty file just before running the transactions.

Why would yum create an empty db file without using it ?

Anyhow somehow yum isn't able to read in the db in /var/lib/rpm
this seems to happen when yum is run and completed,
the pc is left on for another day or two,
and yum can't read in the db anymore.
Delete the db in /var/lib/rpm/__db.* and yum will run again (maybe deleting the empty db /var/lib/rpm/__db.000 will do)
.

[abieleck]

gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-fedora file:///etc/pki/rpm-gpg/RPM-GPG-KEY'

is in a file

/etc/yum.repos.d/fedora-core.repo

I have Fedora Core 6, in other distributions the location may be different.
I had the same problem but typing

rm /var/lib/rpm/__db*
rpm --rebuilddb

was enough to fix the problem. I don't know what this command means, just followed advice.

Regards,
Artur
-----------------------------------------------
A software you don't know how to use is useless

[u-noneinc-s]

There is another thread that explains that this may not alway be the preferred method. I'll post it when I find it, Or do a search for common solutions and you should find it, but I think it says that rpm --rebuilddb could be dangerous and you should do the rm /var/lib/rpm/__db* first and reboot and the --db's should be recreated and the rebuilddb should be a last resort.

I have done this several times (rm --db* and rebuilddb) without issues, but I did it once on FC3 and was royally screwed so I don't doubt that the thread has merits.