[geeks] Package Database Corruption?
velociraptor
velociraptor at gmail.com
Wed Aug 17 19:03:00 CDT 2005
On 8/17/05, James Jacocks <jjacocks at mac.com> wrote:
> I'm trying to replace a package that I created with a new version,
> and I got the following when running pkgrm:
>
> ajacocks at brtbk-64$ sudo pkgrm B2Eossl
>
> The following package is currently installed:
> B2Eossl OpenSSL Libraries
> (sparc) 0.9.7g
>
> Do you want to remove this package? y
>
> ## Removing installed package instance <B2Eossl>
> ## Verifying package dependencies.
> ## Processing package information.
> pkgrm: ERROR: bad read of contents file
> pkgrm: ERROR: pathname=/opt/B2Eperl5/man/man3/overload.3
> pkgrm: ERROR: problem=unable to read content info
>
> That struck me as a bit odd, but I just deleted the presumably bad
> entry in the contents file, and continued. However, the error then
> moved on to another file:
>
> ajacocks at brtbk-64$ sudo pkgrm B2Eossl
>
> The following package is currently installed:
> B2Eossl OpenSSL Libraries
> (sparc) 0.9.7g
>
> Do you want to remove this package? y
> WARNING: Stale lock installed for pkgrm, pkg B2Eossl quit in initial
> state.
> Removing lock.
>
> ## Removing installed package instance <B2Eossl>
> (A previous attempt may have been unsuccessful.)
> ## Verifying package dependencies.
> ## Processing package information.
> pkgrm: ERROR: bad read of contents file
> pkgrm: ERROR: pathname=/usr/dt/appconfig/help/C/graphics/
> TEFileExists.tif
> pkgrm: ERROR: problem=unable to read content info
>
> Anyone here have a suggestion? I'm pretty stumped, beyond the
> obvious system re-install to get the package database back to normal.
>
> Also, does anyone have an idea how this could happen? There hasn't
> been any manual editing of the contents file.
We had a similar problem on a Solaris 8 box that we had
a while back which manifested itself when we tried to
patch the thing (e.g. the patch check scripts claiming
things weren't installed when they were, etc).
Unfortunately, trying to restore the db et al was a no go due
to our backups having some issues at the time. Since it was
a dev box, we just decided to ignore the problem as it
was slated for rebuild.
I did quite a little bit of searching when we discovered the
problem, but didn't find any other suggestions. I figured a
restore was the only possible solution.
=Nadine=
More information about the geeks
mailing list