FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Redhat > Fedora User

 
 
LinkBack Thread Tools
 
Old 12-30-2007, 02:00 AM
Joe Smith
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

Joe Smith wrote:
This is the second update in as many days that has resulted in a
corrupted rpm database for me. ...


Just an update, FWIW.

I continued to have routine rpmdb corruption issues, sometimes with as
simple a command as "yum remove foo-1.2.3-1.fc8". The rebuilddb
procedure was always successful and allowed me to proceed with minimal
package management, but after the next reboot (it's just a desktop:
booted mornings; shut down at night), yum commands would encounter
corruption again.


This continued up until about a week ago, when I decided that I would
try shutting off yum-updatesd altogether, and since then I've had zero
problems with any yum commands, including two updates involving several
packages.


I realize this is purely circumstantial, and I have no explanation why
no one else is reporting trouble, if there's a problem with yum-updatesd.


Since it still wouldn't prove anything to re-enable yum-updatesd and see
the problem return (or not), and I don't care to fiddle with rebuilddb
without a chance to get some solid evidence in return, I'm just going to
leave yum-updatesd off, for now.


If anyone has any thoughts or suggestions for further steps to
understanding this, I'd be most interested to hear them.


<Joe

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 
Old 12-30-2007, 02:14 AM
Craig White
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

On Sat, 2007-12-29 at 22:00 -0500, Joe Smith wrote:
> Joe Smith wrote:
> > This is the second update in as many days that has resulted in a
> > corrupted rpm database for me. ...
>
> Just an update, FWIW.
>
> I continued to have routine rpmdb corruption issues, sometimes with as
> simple a command as "yum remove foo-1.2.3-1.fc8". The rebuilddb
> procedure was always successful and allowed me to proceed with minimal
> package management, but after the next reboot (it's just a desktop:
> booted mornings; shut down at night), yum commands would encounter
> corruption again.
>
> This continued up until about a week ago, when I decided that I would
> try shutting off yum-updatesd altogether, and since then I've had zero
> problems with any yum commands, including two updates involving several
> packages.
>
> I realize this is purely circumstantial, and I have no explanation why
> no one else is reporting trouble, if there's a problem with yum-updatesd.
>
> Since it still wouldn't prove anything to re-enable yum-updatesd and see
> the problem return (or not), and I don't care to fiddle with rebuilddb
> without a chance to get some solid evidence in return, I'm just going to
> leave yum-updatesd off, for now.
>
> If anyone has any thoughts or suggestions for further steps to
> understanding this, I'd be most interested to hear them.
----
I have seen little to no complaints regarding this other than yours. I
suspect that it is not an endemic problem (ala RHL 8.0)

Last time I suggested that you peruse/search bugzilla for other reports
as probably the best way to see if there are reports similar to yours.

Craig

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 
Old 12-30-2007, 09:55 PM
Joe Smith
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

Craig White wrote:
...
Last time I suggested that you peruse/search bugzilla for other reports

as probably the best way to see if there are reports similar to yours.


Ack! You did, and I forgot to look.

So... lots of nasty "rpmdb: PANIC: fatal region error detected..."
reports, mostly fc6 or earlier.


There was one nasty kernel mmap bug, but that was against fc6 and fixed
for kernels >= 2.6.9, although it seems at least one person is still
experiencing it with f7/f8:

https://bugzilla.redhat.com/show_bug.cgi?id=181363

Anyway, it seems that the rpmdb exercises the system in some unusual
ways, and it's not unheard of to see "mysterious" corruption that can
depend on what hardware is present.


I'm just going to stick with manual update checks for now.

<Joe

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 
Old 01-04-2008, 04:15 AM
Joe Smith
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

Joe Smith wrote:

...
I realize this is purely circumstantial, and I have no explanation why
no one else is reporting trouble, if there's a problem with yum-updatesd.


Since it still wouldn't prove anything to re-enable yum-updatesd and see
the problem return (or not), and I don't care to fiddle with rebuilddb
without a chance to get some solid evidence in return, I'm just going to
leave yum-updatesd off, for now.

...


And just now I had the database go belly up even with yum-updatesd
completely disabled, so it looks like that's off the hook.


I wonder how the same hardware was stable all through f7, and f8 up
until about a month ago. Maybe a kernel glitch with my hardware? The
system is stable otherwise, but maybe the database exercises something
unusual.


Still scratching my head...

<Joe

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 
Old 01-04-2008, 04:27 AM
"Kam Leo"
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

On Jan 3, 2008 9:15 PM, Joe Smith <jes@martnet.com> wrote:
> Joe Smith wrote:
> > ...
> > I realize this is purely circumstantial, and I have no explanation why
> > no one else is reporting trouble, if there's a problem with yum-updatesd.
> >
> > Since it still wouldn't prove anything to re-enable yum-updatesd and see
> > the problem return (or not), and I don't care to fiddle with rebuilddb
> > without a chance to get some solid evidence in return, I'm just going to
> > leave yum-updatesd off, for now.
> > ...
>
> And just now I had the database go belly up even with yum-updatesd
> completely disabled, so it looks like that's off the hook.
>
> I wonder how the same hardware was stable all through f7, and f8 up
> until about a month ago. Maybe a kernel glitch with my hardware? The
> system is stable otherwise, but maybe the database exercises something
> unusual.
>
> Still scratching my head...
>
>
> <Joe

There's a chance that your hardware (power supply, fan, RAM, hard
drive, etc.) is failing and you are observing the symptoms.

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 
Old 01-04-2008, 05:01 AM
Craig White
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

On Fri, 2008-01-04 at 00:15 -0500, Joe Smith wrote:
> Joe Smith wrote:
> > ...
> > I realize this is purely circumstantial, and I have no explanation why
> > no one else is reporting trouble, if there's a problem with yum-updatesd.
> >
> > Since it still wouldn't prove anything to re-enable yum-updatesd and see
> > the problem return (or not), and I don't care to fiddle with rebuilddb
> > without a chance to get some solid evidence in return, I'm just going to
> > leave yum-updatesd off, for now.
> > ...
>
> And just now I had the database go belly up even with yum-updatesd
> completely disabled, so it looks like that's off the hook.
>
> I wonder how the same hardware was stable all through f7, and f8 up
> until about a month ago. Maybe a kernel glitch with my hardware? The
> system is stable otherwise, but maybe the database exercises something
> unusual.
----
almost sounds like an incomplete upgrade occurred, perhaps with the
wrong version of berkley-db4 or python installed.

Did you ever go through things like installing yum-tools and
package-cleanup?

Craig

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 
Old 01-04-2008, 04:17 PM
Joe Smith
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

Craig White wrote:

...
almost sounds like an incomplete upgrade occurred, perhaps with the
wrong version of berkley-db4 or python installed.

Did you ever go through things like installing yum-tools and
package-cleanup?


Hey, thanks for the ideas!

"rpm -V db4" was clean--something else I should check?

I always assumed that if "rpm --rebuilddb" succeeded, then the database
should be ok. But it looks like my assumption was wrong:


# cd /var/lib/rpm/
# /usr/lib/rpm/rpmdb_verify Packages
# /usr/lib/rpm/rpmdb_verify Requirename
db_verify: Page 223: bad prev_pgno 0 on overflow page (should be 119)
db_verify: Page 223: overflow item incomplete
db_verify: Requirename: DB_VERIFY_BAD: Database verification failed
# for f in *; do echo "$f"; /usr/lib/rpm/rpmdb_verify $f; done
Basenames
Conflictname
Dirnames
db_verify: Page 1982: bad page number 191
db_verify: Dirnames: DB_VERIFY_BAD: Database verification failed
Filemd5s
db_verify: Page 4635: bad page number 89
db_verify: Filemd5s: DB_VERIFY_BAD: Database verification failed
Group
db_verify: Page 18: overflow page of invalid type 2
db_verify: Group: DB_VERIFY_BAD: Database verification failed
Installtid
Name
Packages
Providename
db_verify: Page 342: bad page number 105
db_verify: Providename: DB_VERIFY_BAD: Database verification failed
Provideversion
db_verify: Page 110: bad prev_pgno 0 on overflow page (should be 112)
db_verify: Page 110: overflow item incomplete
db_verify: Provideversion: DB_VERIFY_BAD: Database verification failed
Pubkeys
Requirename
db_verify: Page 223: bad prev_pgno 0 on overflow page (should be 119)
db_verify: Page 223: overflow item incomplete
db_verify: Requirename: DB_VERIFY_BAD: Database verification failed
Requireversion
db_verify: Page 191: partially zeroed page
db_verify: Requireversion: DB_VERIFY_BAD: Database verification failed
Sha1header
Sigmd5
db_verify: Page 32: bad page number 223
db_verify: Page 32: invalid prev_pgno 119
db_verify: Sigmd5: DB_VERIFY_BAD: Database verification failed
Triggername

I found an old (2003) suggestion for rebuilding the Packages db:
> mv Packages Packages-ORIG
> /usr/lib/rpm/rpmdb_dump Packages-ORIG |
> /usr/lib/rpm/rpmdb_load Packages

Running that on each db...
# for f in Requireversion Provideversion Providename Group Filemd5s
Dirnames; do echo "$f"; mv "$f" "$f.orig" && /usr/lib/rpm/rpmdb_dump
"$f.orig" | /usr/lib/rpm/rpmdb_load "$f"; done

Requireversion
db_load: Transactional Data Store incompatible with environment
Provideversion
db_load: Transactional Data Store incompatible with environment
Providename
db_load: Transactional Data Store incompatible with environment
Group
db_load: Transactional Data Store incompatible with environment
Filemd5s
db_load: Transactional Data Store incompatible with environment
Dirnames
db_load: Transactional Data Store incompatible with environment

I have no idea what that message means, but it doesn't mean the command
failed, apparently:


# for f in `ls | grep -v '.'`; do echo "$f"; /usr/lib/rpm/rpmdb_verify
$f; done

Basenames
Conflictname
Dirnames
Filemd5s
Group
Installtid
Name
Packages
Providename
Provideversion
Pubkeys
Requirename
Requireversion
Sha1header
Sigmd5
Triggername

Looks clean now.

I'm running a "rpm -Va" just now, and so far it looks reasonable, so
maybe the problem is done now.


I have no idea what I'm doing here, and I still don't understand if
"--rebuilddb" was missing something, or what.


Clues welcomed.

PS: Oops. No, now the "rpm -Va" is finding stuff that looks like some
parts of the last update that failed were partially installed. Maybe I
need to go back and repeat the latest package installs to make sure
everything is clean.


Maybe time to punt and re-install f8.

<Joe

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 
Old 01-04-2008, 04:27 PM
Craig White
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

On Fri, 2008-01-04 at 12:17 -0500, Joe Smith wrote:
> Craig White wrote:
> > ...
> > almost sounds like an incomplete upgrade occurred, perhaps with the
> > wrong version of berkley-db4 or python installed.
> >
> > Did you ever go through things like installing yum-tools and
> > package-cleanup?
>
> Hey, thanks for the ideas!
>
> "rpm -V db4" was clean--something else I should check?
>
> I always assumed that if "rpm --rebuilddb" succeeded, then the database
> should be ok. But it looks like my assumption was wrong:
>
> # cd /var/lib/rpm/
> # /usr/lib/rpm/rpmdb_verify Packages
> # /usr/lib/rpm/rpmdb_verify Requirename
> db_verify: Page 223: bad prev_pgno 0 on overflow page (should be 119)
> db_verify: Page 223: overflow item incomplete
> db_verify: Requirename: DB_VERIFY_BAD: Database verification failed
> # for f in *; do echo "$f"; /usr/lib/rpm/rpmdb_verify $f; done
> Basenames
> Conflictname
> Dirnames
> db_verify: Page 1982: bad page number 191
> db_verify: Dirnames: DB_VERIFY_BAD: Database verification failed
> Filemd5s
> db_verify: Page 4635: bad page number 89
> db_verify: Filemd5s: DB_VERIFY_BAD: Database verification failed
> Group
> db_verify: Page 18: overflow page of invalid type 2
> db_verify: Group: DB_VERIFY_BAD: Database verification failed
> Installtid
> Name
> Packages
> Providename
> db_verify: Page 342: bad page number 105
> db_verify: Providename: DB_VERIFY_BAD: Database verification failed
> Provideversion
> db_verify: Page 110: bad prev_pgno 0 on overflow page (should be 112)
> db_verify: Page 110: overflow item incomplete
> db_verify: Provideversion: DB_VERIFY_BAD: Database verification failed
> Pubkeys
> Requirename
> db_verify: Page 223: bad prev_pgno 0 on overflow page (should be 119)
> db_verify: Page 223: overflow item incomplete
> db_verify: Requirename: DB_VERIFY_BAD: Database verification failed
> Requireversion
> db_verify: Page 191: partially zeroed page
> db_verify: Requireversion: DB_VERIFY_BAD: Database verification failed
> Sha1header
> Sigmd5
> db_verify: Page 32: bad page number 223
> db_verify: Page 32: invalid prev_pgno 119
> db_verify: Sigmd5: DB_VERIFY_BAD: Database verification failed
> Triggername
>
> I found an old (2003) suggestion for rebuilding the Packages db:
> > mv Packages Packages-ORIG
> > /usr/lib/rpm/rpmdb_dump Packages-ORIG |
> > /usr/lib/rpm/rpmdb_load Packages
>
> Running that on each db...
> # for f in Requireversion Provideversion Providename Group Filemd5s
> Dirnames; do echo "$f"; mv "$f" "$f.orig" && /usr/lib/rpm/rpmdb_dump
> "$f.orig" | /usr/lib/rpm/rpmdb_load "$f"; done
> Requireversion
> db_load: Transactional Data Store incompatible with environment
> Provideversion
> db_load: Transactional Data Store incompatible with environment
> Providename
> db_load: Transactional Data Store incompatible with environment
> Group
> db_load: Transactional Data Store incompatible with environment
> Filemd5s
> db_load: Transactional Data Store incompatible with environment
> Dirnames
> db_load: Transactional Data Store incompatible with environment
>
> I have no idea what that message means, but it doesn't mean the command
> failed, apparently:
>
> # for f in `ls | grep -v '.'`; do echo "$f"; /usr/lib/rpm/rpmdb_verify
> $f; done
> Basenames
> Conflictname
> Dirnames
> Filemd5s
> Group
> Installtid
> Name
> Packages
> Providename
> Provideversion
> Pubkeys
> Requirename
> Requireversion
> Sha1header
> Sigmd5
> Triggername
>
> Looks clean now.
>
> I'm running a "rpm -Va" just now, and so far it looks reasonable, so
> maybe the problem is done now.
>
> I have no idea what I'm doing here, and I still don't understand if
> "--rebuilddb" was missing something, or what.
>
> Clues welcomed.
>
> PS: Oops. No, now the "rpm -Va" is finding stuff that looks like some
> parts of the last update that failed were partially installed. Maybe I
> need to go back and repeat the latest package installs to make sure
> everything is clean.
>
> Maybe time to punt and re-install f8.
----
see this...

http://fedoraproject.org/wiki/YumUpgradeFaq

see point #4 'Tips' about yum-utils/package-cleanup

Craig

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 
Old 01-05-2008, 05:07 AM
Joe Smith
 
Default maybe yum-updatesd? (was yum updates corrupting rpmdb)

Joe Smith wrote:

...
I'm running a "rpm -Va" just now, and so far it looks reasonable, so
maybe the problem is done now.

...


Nope. Yum can even successfully finish an update which apparently leaves
errors in the database:


# yum update
Setting up Update Process
Resolving Dependencies
--> Running transaction check
---> Package gnupg2.i386 0:2.0.8-1.fc8 set to be updated
---> Package gdb.i386 0:6.6-39.fc8 set to be updated
---> Package liferea.i386 0:1.4.10-1.fc8 set to be updated
---> Package libEMF.i386 0:1.0.3-5.fc8 set to be updated
---> Package yum-utils.noarch 0:1.1.9-2.fc8 set to be updated
---> Package libcdio.i386 0:0.78.2-4.fc8 set to be updated
--> Finished Dependency Resolution
...
Updated: gdb.i386 0:6.6-39.fc8 gnupg2.i386 0:2.0.8-1.fc8 libEMF.i386
0:1.0.3-5.fc8 libcdio.i386 0:0.78.2-4.fc8 liferea.i386 0:1.4.10-1.fc8
yum-utils.noarch 0:1.1.9-2.fc8

Complete!
# cd /var/lib/rpm
# for f in `ls | grep -v '.'`; do /usr/lib/rpm/rpmdb_verify $f; done
db_verify: Page 2355: partially zeroed page
db_verify: Dirnames: DB_VERIFY_BAD: Database verification failed
db_verify: Page 641: bad page number 247
db_verify: Providename: DB_VERIFY_BAD: Database verification failed
db_verify: Page 109: partially zeroed page
db_verify: Provideversion: DB_VERIFY_BAD: Database verification failed
db_verify: Page 247: partially zeroed page
db_verify: Requireversion: DB_VERIFY_BAD: Database verification failed
db_verify: Page 160: bad page number 109
db_verify: Sha1header: DB_VERIFY_BAD: Database verification failed

If it hadn't ever worked, I'd be looking at hardware, but it's worked
well for many months (f7/8), and I'm seeing no drive errors or
instability otherwise.


<Joe

--
fedora-list mailing list
fedora-list@redhat.com
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
 

Thread Tools




All times are GMT. The time now is 03:12 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org