Re: RPM problems with RedHat8.0

this is a known problem with the rpm in redhat 8
a reboot is not necessary...and the problem is inconsistant at best..and
magically delicious at worst.

there were a set of testing rpms by a redhat developer out and about
that MIGHT fix this problem...and I am running the rawhide rpm versions
as they come out in the hopes of having this problem go away. It doesnt
happen often for me at this point, but it does happen still....not a
grab problem....but grab might be doing something particularly
interesting enough to make the problem show up....all I know is that I
can't consistantly reproduce the problem...so its a pain in the ass to
try to explore.  It happens when it happens....but it happens.

the only known workaround is to rm the __db* files in /var/lib/rpm

Let's see if I can dig up the bugzilla ticket(s) for this problem. One
of the bugtixs had the ftp location for the testing rpms.

prefixing the rm __db* command to certain things that grab does is prob
not a bad workaround for grab....but i cant get grab to consistantly
produce the problem...so anything you do at this point would be voodoo
magic and not very scientific.

-jef

On Fri, 2002-11-22 at 20:33, Greg Kurtzer wrote:
> At my work there have been several reports of RPM just locking and not
> working. It was thought even that GRAB may be responsible for this. I finally
> had the chance to hack on one of these systems, and it turns out that there
> were some stale/corrupt database files that were the culprit. GRAB was only
> involved because it was what was calling RPM to do the install.
> 
> If you have any problems with RPM 4.1 locking (seemingly) beyond repair, try 
> to remove /var/lib/rpm/__db.* and try again. These files seem to be temporary
> cache files and it will not harm your RPM database to remove them.
> 
> I am not sure what is causing this, and I personally have never seen the
> problem on any of my systems. On the RPM mailing list it mentioned that
> similar problems were caused by killing RPM in the middle of a job. But the 
> admins were telling me that they only killed the command after it hung for 10+ 
> minutes. Anyway, after removing these stale files, RPM was able to continue 
> normally.
> 
> Interesting side-note:
> 
> [gmkurtzer@spike ~]$ grep __db /etc/rc.sysinit
> rm -f /var/lib/rpm/__db*
> 
> So I guess this is one area that RedHat has taken after Windows. Rebooting
> _should_ fix this problem! That just pains me to say. Hey Jeff, if you are
> listening, maybe this command should just be prefixed in /usr/lib/rpm/rpmd
> rebuilddb sequence?
> 
> Greg

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQA93s7srWLDmRitRZURAvR9AJ4mU9QW9W6c4WQPx2HdhmpzD8MdFgCaAj7M
uLBap7w2hs9F5PgrbcazmEA=
=4ONo
-----END PGP SIGNATURE-----