[SOLVED] HELP-I've broken rpm

[jaywalker]

First I installed mplayer-gui along with another batch of updates.
Then I tried to run it and then I installed the 1.2 version of libpng.
Then I changed my mind and tried to uninstall it but rpmdrake wouldn't open - failed with a database error message.
So then I tried this:

Code: Select all

[root@Tureen rich]# urpme libpng
error: rpmdb: Thread/process 2573/139782153705216 failed: Thread died in Berkeley DB library
error: db4 error(-30974) from dbenv->failchk: DB_RUNRECOVERY: Fatal error, run database recovery
error: cannot open Packages index using db4 -  (-30974)
error: cannot open Packages database in /var/lib/rpm
unable to open rpmdb


and then this:

Code: Select all

[root@Tureen rich]# rpmdb --rebuilddb
error: rpmdb: Thread/process 2573/139782153705216 failed: Thread died in Berkeley DB library
error: db4 error(-30974) from dbenv->failchk: DB_RUNRECOVERY: Fatal error, run database recovery
error: cannot open Packages index using db4 -  (-30974)


and finally this;

Code: Select all

[root@Tureen rich]# rpmdb --initdb
error: rpmdb: Thread/process 2573/139782153705216 failed: Thread died in Berkeley DB library
error: db4 error(-30974) from dbenv->failchk: DB_RUNRECOVERY: Fatal error, run database recovery
error: cannot open Packages index using db4 -  (-30974)


What is a Berkeley DB and what tools do I have to kick it back into shape? All I have is what I have, though I could probably download and manually install anything I might be missing.

All suggestions gratefully received...

Richard

[doktor5000]

You might want to have a look at: http://www.oldrpm.org/hintskinks/repairdb/
Berkeley DB is the database RPM uses to store metadata, look at some of the mails at the bottom to know how to manually check some parts of that database.

[jaywalker]

Thanks Dok, past my bedtime now - I'll explore that tomorrow evening.

Richard

[jaywalker]

OK, so curiosity got the better of me and now everything is working just fine. I did a

Code: Select all

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

to get rid of 5 files; one dating from 3 March (zero length) and four from tonight (very large).

I don't understand why it works, so that will have to wait for more research later.

Thanks again

Richard

[DShelbyD]

Roughly the same issue here today. Thanks to both of you, things appear to be OK now. A couple of "bad" and/or "missing" signatures, but I went ahead anyway. So far, so good.

Code: Select all

The following packages have bad signatures:                                   
/var/cache/urpmi/rpms/lib64rpmbuild2-4.9.1.2-21.mga2.x86_64.rpm: Missing signature (OK ((none)))
/var/cache/urpmi/rpms/lib64rpmsign2-4.9.1.2-21.mga2.x86_64.rpm: Missing signature (OK ((none)))

[ozky]

I have same problem too it broke rpm database and then urpmi --auto-select is still says that need to install rpm update and it fail that install because it is already installed,and it will try to install it again and again,i have rebuild rpm database but it will give no help to this problem.

[jaywalker]

And my problem came back a few minutes ago after installing this evening's updates - mostly rpm and file system layout stuff, from what I recall - first the rpm updates went OK, then the restart and display of half a dozen more. rpmdrake crashed during the installation of these last few.

[alf]

I was also trapped this morning after today's updates. The rpm rebuilddb didn't help so i removed all sources and added them newly. Now it seems to work.

[DShelbyD]

FWIW I'm still OK after the updates of late 13 March (new rpm + others).

[jaywalker]

Despite the two rpm updates we had today, 4.9.1.2-21.mga2 and 4.9.1.2-22.mga2, I am still experiencing the problem. It is under investigation as https://bugs.mageia.org/show_bug.cgi?id=4918.

It can still be temporarily cleared by;

Code: Select all

 rm -f /var/lib/rpm/_*
rpmd --rebuilddb

I also delete /var/lib/rpm/.RPMLOCK, but I don't know if it makes a difference.

[DShelbyD]

Yes, after applying the mageia version update and one other package this morning, the previous DB problem returned. However, I applied

Code: Select all

rm -f /var/lib/rpm/_*

and then did

Code: Select all

urpmi --auto-update

Here is the session:

Code: Select all

[root@Athlon6000 ~]# urpmi --auto-update
medium "Core Release (distrib1)" is up-to-date
medium "Core Updates (distrib3)" is up-to-date
medium "Nonfree Release (distrib11)" is up-to-date
medium "Nonfree Updates (distrib13)" is up-to-date
medium "Tainted Release (distrib21)" is up-to-date
medium "Tainted Updates (distrib23)" is up-to-date
medium "Core 32bit Release (distrib31)" is up-to-date
medium "Core 32bit Updates (distrib33)" is up-to-date
To satisfy dependencies, the following package is going to be installed:
   Package                        Version      Release       Arch   
(medium "Core Release (distrib1)")
  meta-task                      2            33.mga2       noarch 
127B of additional disk space will be used.
33KB of packages will be retrieved.
Proceed with the installation of one package? (Y/n)


    ftp://mirrors.kernel.org/mageia/distrib/cauldron/x86_64/media/core/release/meta-task-2-33.mga2.noarch.rpm
installing meta-task-2-33.mga2.noarch.rpm from /var/cache/urpmi/rpms           
Preparing...                     #############################################
      1/1: meta-task             #############################################
restarting urpmi
To satisfy dependencies, the following packages are going to be installed:
   Package                        Version      Release       Arch   
(medium "Core Release (distrib1)")
  dhcp-client                    4.2.3P2      4.mga2        x86_64 
  dhcp-common                    4.2.3P2      4.mga2        x86_64 
  mageia-release-Default         2            0.7.mga2      x86_64 
0B of additional disk space will be used.
684KB of packages will be retrieved.
Proceed with the installation of the 3 packages? (Y/n)


    ftp://mirrors.kernel.org/mageia/distrib/cauldron/x86_64/media/core/release/dhcp-client-4.2.3P2-4.mga2.x86_64.rpm
    ftp://mirrors.kernel.org/mageia/distrib/cauldron/x86_64/media/core/release/dhcp-common-4.2.3P2-4.mga2.x86_64.rpm
                                                                               
installing dhcp-client-4.2.3P2-4.mga2.x86_64.rpm dhcp-common-4.2.3P2-4.mga2.x86_64.rpm mageia-release-Default-2-0.7.mga2.x86_64.rpm from /var/cache/urpmi/rpms
Preparing...                     #############################################
      1/3: dhcp-common           #############################################
      2/3: dhcp-client           #############################################
      3/3: mageia-release-Default
                                 #############################################
[root@Athlon6000 ~]#

Only FWIW,

[wobo]

Yes, after applying the mageia version update and one other package this morning, the previous DB problem returned. However, I applied

Code: Select all

rm -f /var/lib/rpm/_*

and then did

Code: Select all

urpmi --auto-update

Does not work here.

1 hour ago I did the latest updates with urpmi, no problems. Then I opened rpmdrake, marking all -handbook packages for uninstall. Uninstalling worked for 2 packages then I faced a segmentation fault.
Opening rpmdrake again showed the known error and advice to rebuild the database.
I removed _db* and did 'rpm --rebuilddb' but this returned:

Code: Select all

Error: rpmdb: Thread/process 2794/1397..... failed: Thread died in Berkeley DB library
Error: db4 error from dbenv->failchk: DB_RUNRECOVERY: Fatal error, run database recovery
Error: Unable to open packages index, using db4

Result: system is unusable, I can't even update or whatever.

Solution:
I erased the _db* files in /var/lib/rpm via and removed the rpm lock files. After that I could rebuild the rpm database.

[wobo]

Hours later and after doing the rebuild dance 4 times more I can say:

1. System is updated to latest status as of 17:01 Paris time
2. It is not a rpm nor urpmi issue. Both work fine.
3. It is an rpmdrake issue - removing any installed package with rpmdrake results in a segmentation fault. Reproducible with any installed package. Uninstalling the same package with urpme works fine.
4. the error message about not being able to access the database is a result of the segfault (rpmdrake leaves it locked and corrupted).

[pmithrandir]

I get that

Code: Select all

[root@portablePierre ~]# urpmi --auto-update
    ftp://ftp.mandrivauser.de/mirrors/Mageia/distrib/cauldron/i586/media/core/release/media_info/20120314-190816-synthesis.hdlist.cz                           
média « Core Release (distrib1) » mis à jour                                   
le média « Core Updates (distrib3) » est à jour                                 
le média « Core Backports (distrib7) » est à jour                               
le média « Nonfree Release (distrib11) » est à jour                             
le média « Nonfree Updates (distrib13) » est à jour                             
le média « Nonfree Backports (distrib17) » est à jour                           
le média « Tainted Release (distrib21) » est à jour                             
le média « Tainted Updates (distrib23) » est à jour                             
le média « Tainted Backports (distrib27) » est à jour                           
Afin de poursuivre la mise à jour, les paquetages suivants doivent être désinstallés :                                                                         
meta-task-2-33.mga2.noarch                                                     
 (pour installer le paquetage meta-task-2-33.mga2.noarch)                       
rpm-4.9.1.2-22.mga2.i586                                                       
 (pour installer le paquetage rpm-4.9.1.2-22.mga2.i586) (o/N) o                 
Pour satisfaire les dépendances, les paquetages suivants vont être installés :
   Paquetage                      Version      Révision      Arch   
(média « Core Release (distrib1) »)
  meta-task                      2            33.mga2       noarch 
  rpm                            4.9.1.2      22.mga2       i586   
un espace de 1.6Mo sera libéré.
447Ko de paquets seront récupérés.
Procéder à l'installation des 2 paquetages ? (O/n) o


installation de meta-task-2-33.mga2.noarch.rpm rpm-4.9.1.2-22.mga2.i586.rpm depuis /var/cache/urpmi/rpms
Préparation ...                  #############################################
L'installation a échoué :       paquetage rpm-1:4.9.1.2-22.mga2.i586 déjà installé
        paquetage meta-task-1:2-33.mga2.noarch déjà installé



any ideas how to resolve it ?

I did the rm command, and the build db one also...

[jaywalker]

https://bugs.mageia.org/show_bug.cgi?id=4918

According to the bug history it looks like this one is getting a lot of attention.

3. It is an rpmdrake issue - removing any installed package with rpmdrake results in a segmentation fault. Reproducible with any installed package. Uninstalling the same package with urpme works fine.4. the error message about not being able to access the database is a result of the segfault (rpmdrake leaves it locked and corrupted).

My experience this evening seems to support wobo's findings so you could try just using urpmi and friends. When you get another update flag, just ignore it and execute the update in a console. If that keeps you clear of the problem then it would be worth adding that information to the bug tracker.

Richard

[pmithrandir]

what else can do in console?
thanks

[wilcal]

This worked for me:

rm -f /var/lib/rpm/_*
rpmd --rebuilddb

[jaywalker]

I know it sounds stupid but do you think it might help to repeat the rpm tidy-up and rebuild the db AGAIN. You might also clean up in /var/cache/urpmi/rpms. It can't do any harm; at worst you will just download some packages again.

I have been deleting /var/lib/__db00*, /var/lib/rpm/.RPMLOCK and /var/lib/urpmi/.LOCK files so often tonight I am seriously considering putting it all in a script (not forgetting rpmdb --rebuilddb).

[wobo]

I have now tested for a while in various configurations of Mageia Cauldron:
1. From a clean system urpm* and rpm are working as they should.
2. After latest updates I had no problems with rpmdrake any more.
rpm-4.9.1.2-22.mga2
rpmdrake-5.30-mga2

Problems encountered before:
1. Using the GUI rpmdrake (either starting in a konsole or via MCC) for uninstalling or installing a package results in a segmentation fault (window closes without error message, if started from the konsole window you see the message "segmentation fault")
2. Restarting rpmdrake or trying to use urpm* results in the error messages about not being able to use the database.

For temporary solution I did:
1. Go to konsole window, become root
2. Do these commands one after the other:

Code: Select all

rm -f /var/lib/rpm/_db*
rm -f /var/lib/rpm/.RPMLOCK
rpm --rebuilddb

After this your urpm* and rpm should be working as normal.

[DShelbyD]

After latest updates I had no problems with rpmdrake any more.
rpm-4.9.1.2-22.mga2
rpmdrake-5.30-mga2

rpm - yes, that is the version I have. But

Code: Select all

[root@Athlon6000 ~]# rpm -q rpmdrake
rpmdrake-5.30-1.mga2

No, wait. As I type another rpmdrake update to 5.30-2.mga2 followed by the return of the familiar index/DB issues. Now for me rpmdrake fails, as does urpmi --auto-update.

The rpm --rebuilddb command, issued following the other two, results in

Code: Select all

error: rpmdb: Thread/process 8567/140113150666496 failed: Thread died in Berkeley DB library
error: db4 error(-30974) from dbenv->failchk: DB_RUNRECOVERY: Fatal error, run database recovery
error: cannot open Packages index using db4 -  (-30974)

[EDIT]In addition, and to further complicate things,

Code: Select all

[root@Athlon6000 ~]# rpm -q rpm
error: rpmdb: Thread/process 8567/140113150666496 failed: Thread died in Berkeley DB library
error: db4 error(-30974) from dbenv->failchk: DB_RUNRECOVERY: Fatal error, run database recovery
error: cannot open Packages index using db4 -  (-30974)
error: cannot open Packages database in /var/lib/rpm
error: rpmdb: Thread/process 8567/140113150666496 failed: Thread died in Berkeley DB library
error: db4 error(-30974) from dbenv->failchk: DB_RUNRECOVERY: Fatal error, run database recovery
error: cannot open Packages database in /var/lib/rpm
package rpm is not installed

And it cannot be installed under the circumstances.

[EDIT again] I solved the issue from the /var/lib/rpm directory with

Code: Select all

rm -f __db*
rm -f .rpm.lock
rm -f rebuilddb

Note in the first: two underscore characters instead of one
Note in the second: lower case and a dot separating "rpm" from "lock"
Now I can urpmi successfully, and rpm shows up as present and current.

Code: Select all

[root@Athlon6000 rpm]# rpm -q rpm
rpm-4.9.1.2-22.mga2

[wobo]

[EDIT again] I solved the issue from the /var/lib/rpm directory with

Code: Select all

rpm -f __db*
rpm -f .rpm.lock
rpm -f rebuilddb

Note in the first: two underscore characters instead of one
Note in the second: lower case and a dot separating "rpm" from "lock"
Now I can urpmi successfully, and rpm shows up as present and current.

Code: Select all

[root@Athlon6000 rpm]# rpm -q rpm
rpm-4.9.1.2-22.mga2

Ah, ok. The missing underscore was a typo, sorry about that.
As for the lockfile: At first I had 2 of them, one with the capital letters and the other as you describe. At the next test and all others I only saw the file with the capital letters. Sorry about forgetting that.

[DShelbyD]

Ego te absolvo!

I think this problem is fixed with the new RPM. Do you? Is anyone still having problems?

[wobo]

Can say late tonight, today my 2 machines with Cauldron are sleeping at home and I am still sitting outside with my laptop, nice spring weather!

[DShelbyD]

Only one machine have I running Cauldron, and it's doing very well now, having just run rpmdrake for a few library updates. The spring weather here in the Mid-Atlantic of the U.S. is downright radiant with trees and flowers bursting. Great day. I am loathe, however, to start a "How's the weather" thread!! Spare me. Right now I'm filled with gratitude.

[pmithrandir]

I'm going to check, but last time I tried I couldn't perform the meta task installation, and there was no other updates. I will check this evening if I have time before my week end.