missing RPM in desktop install -- xsri-2.1.0-6

[Alan Peery]

I get 80%+ through installation of a default desktop installation, and 
the installation stops at "The package xsri-2.1.0-6 cammnot be opened.  
This is due to a missing file or perhaps a corrup package..."

At this point the install will let me retry forever, but not skip the 
package and continue on.  Is there any way to work around this?  The 
md5sums of the .iso files are correct, so I don't think I've had a 
download problem.

Detail:
I happen to be installing the software under VMWare (version 3.0, 
running on a Windows XP host).  I've swapped the .iso files by 
suspending the virtual machine and "swapping cdroms" by renaming the 
.iso files on my hard drive.  I've used this technique many times, and 
it worked for 20 minutes following the switch, so I don't think this is 
likely to be the problem.

Alan

[Jack Aboutboul]

On Wednesday, Oct 22, 2003, at 14:18 America/New_York, Alan Peery wrote:
> I get 80%+ through installation of a default desktop installation, and 
> the installation stops at "The package xsri-2.1.0-6 cammnot be opened. 
>  This is due to a missing file or perhaps a corrup package..."
>
> At this point the install will let me retry forever, but not skip the 
> package and continue on.  Is there any way to work around this?  The 
> md5sums of the .iso files are correct, so I don't think I've had a 
> download problem.
>
> Detail:
> I happen to be installing the software under VMWare (version 3.0, 
> running on a Windows XP host).  I've swapped the .iso files by 
> suspending the virtual machine and "swapping cdroms" by renaming the 
> .iso files on my hard drive.  I've used this technique many times, and 
> it worked for 20 minutes following the switch, so I don't think this 
> is likely to be the problem.
>
> Alan

Alan,

I've had this problem too, and never got it figured out. I think its 
because you need to have vmware-tools installed before you can install 
xsri, however, thats not possible to install unless the dist install is 
complete. Officially, installing under vmware is not supported. Please 
keep me updated as this is a bug I would love to squash.

Thanks,
Jack

[Mike A. Harris]

On Wed, 22 Oct 2003, Jack Aboutboul wrote:
> > I get 80%+ through installation of a default desktop installation, and 
> > the installation stops at "The package xsri-2.1.0-6 cammnot be opened. 
> >  This is due to a missing file or perhaps a corrup package..."
> >
> > At this point the install will let me retry forever, but not skip the 
> > package and continue on.  Is there any way to work around this?  The 
> > md5sums of the .iso files are correct, so I don't think I've had a 
> > download problem.
> >
> > Detail:
> > I happen to be installing the software under VMWare (version 3.0, 
> > running on a Windows XP host).  I've swapped the .iso files by 
> > suspending the virtual machine and "swapping cdroms" by renaming the 
> > .iso files on my hard drive.  I've used this technique many times, and 
> > it worked for 20 minutes following the switch, so I don't think this 
> > is likely to be the problem.
> >
> > Alan
>
> Alan,
>
> I've had this problem too, and never got it figured out. I think its 
> because you need to have vmware-tools installed before you can install 
> xsri, however, thats not possible to install unless the dist install is 
> complete. Officially, installing under vmware is not supported. Please 
> keep me updated as this is a bug I would love to squash.

I suspect the media is bad.  Please run mediacheck and test your 
media first.  I don't see why this problem would occur only under 
VMware personally.  Even though we don't support VMware 
officially, problems like this still shouldn't happen.

[Mike - W0TMW]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wednesday 22 October 2003 01:18 pm, Alan Peery wrote:
> I get 80%+ through installation of a default desktop installation,
> and the installation stops at "The package xsri-2.1.0-6 cammnot be
> opened. This is due to a missing file or perhaps a corrup package..."
>
> At this point the install will let me retry forever, but not skip the
> package and continue on.  Is there any way to work around this?  The
> md5sums of the .iso files are correct, so I don't think I've had a
> download problem.
>
> Detail:
> I happen to be installing the software under VMWare (version 3.0,
> running on a Windows XP host).  I've swapped the .iso files by
> suspending the virtual machine and "swapping cdroms" by renaming the
> .iso files on my hard drive.  I've used this technique many times,
> and it worked for 20 minutes following the switch, so I don't think
> this is likely to be the problem.
>
> Alan

Are you installing off CDs?  If so, run mediacheck.  I've downloaded 
ISOs, had MD5SUM work OK, but had a bad write to the CD and 
subsequently fail.

Mike W
- -- 
Registered Linux - 256979   (http://counter.il.org for more information)
NRA Life
ARS: W0TMW
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.7 (GNU/Linux)
>
> iD8DBQE/lyKT5fq6h2uDDlQRAmLwAJ4uU20kk1UXAja3+Lp8NBj+2ZRwbgCgqSQH
> S2ZrFRjUNIPqVy8OB+OZ+Lg=
> =r5LL
> -----END PGP SIGNATURE-----

[Duncan Morison]

Never ever succesfully installed redhat on vmware via
the CDs... this has always happenned to me.

But if you select the CD drive to mount the redhat
iso's it will work fine.

used to happen on vmware 3 and still happens with 4.

not quite sure why this is the case but it seems to
happen to many people who have tried to install rh on
vmware from what I've seen.

seeing as mounting the iso's always works 'shrug'

--- Mike Watson  wrote: >
-----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Wednesday 22 October 2003 01:18 pm, Alan Peery
> wrote:
> > I get 80%+ through installation of a default
>
> desktop installation,
> > and the installation stops at "The package
>
> xsri-2.1.0-6 cammnot be
> > opened. This is due to a missing file or perhaps a
>
> corrup package..."
> > At this point the install will let me retry
>
> forever, but not skip the
> > package and continue on.  Is there any way to work
>
> around this?  The
> > md5sums of the .iso files are correct, so I don't
>
> think I've had a
> > download problem.
> >
> > Detail:
> > I happen to be installing the software under
>
> VMWare (version 3.0,
> > running on a Windows XP host).  I've swapped the
>
> .iso files by
> > suspending the virtual machine and "swapping
>
> cdroms" by renaming the
> > .iso files on my hard drive.  I've used this
>
> technique many times,
> > and it worked for 20 minutes following the switch,
>
> so I don't think
> > this is likely to be the problem.
> >
> > Alan
>
> Are you installing off CDs?  If so, run mediacheck. 
> I've downloaded 
> ISOs, had MD5SUM work OK, but had a bad write to the
> CD and 
> subsequently fail.
>
> Mike W
>
> Registered Linux - 256979   (http://counter.il.org
> for more information)
> NRA Life
> ARS: W0TMW
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.0.7 (GNU/Linux)

iD8DBQE/lyKT5fq6h2uDDlQRAmLwAJ4uU20kk1UXAja3+Lp8NBj+2ZRwbgCgqSQH
> S2ZrFRjUNIPqVy8OB+OZ+Lg=
> =r5LL
> -----END PGP SIGNATURE-----
>
> -- 
> This message has been scanned for viruses and 
> dangerous content by MailScanner, 
> and is believed to be clean.
>
> --
> fedora-list mailing list
> [e-mail]
> http://www.redhat.com/mailman/listinfo/fedora-list

________________________________________________________________________
Want to chat instantly with your online friends?  Get the FREE Yahoo!
Messenger http://mail.messenger.yahoo.co.uk

[Chris Spencer]

On Wed, 2003-10-22 at 15:52, Mike A. Harris wrote:
> > > I get 80%+ through installation of a default desktop installation, and 
> > > the installation stops at "The package xsri-2.1.0-6 cammnot be opened. 
> > >  This is due to a missing file or perhaps a corrup package..."

I see this all the time.  Why is it that the installer doesn't allow you
to retry a package ... and instead errors out and dies...

I actually tested my media and I still had the problem at some point in
my installation.

Bottom line though it shouldn't just die.  Give a chance to retry,
abort, or ignore...and continue giving that chance until the USER
decides to abort.

Maybe I have 3 or for disks...1 may go bad but I have more.  I can burn
new ones in a couple of minutes...etc.

-Chris

"Restriction of free thought and free speech is the most dangerous of
all subversions. It is the one un-American act that could most easily
defeat us." - Supreme Court Justice William O. Douglas

[Niels Weber]

Am Do, den 23.10.2003 schrieb Chris Spencer um 17:01:
> On Wed, 2003-10-22 at 15:52, Mike A. Harris wrote:
> > > > I get 80%+ through installation of a default desktop installation, and 
> > > > the installation stops at "The package xsri-2.1.0-6 cammnot be opened. 
> > > >  This is due to a missing file or perhaps a corrup package..."
>
> I see this all the time.  Why is it that the installer doesn't allow you
> to retry a package ... and instead errors out and dies...

I even had the problem that I couldn't get my CD out anymore. I had to
completly switch of the computer for that.

> I see this all the time.  Why is it that the installer doesn't allow you
> to retry a package ... and instead errors out and dies...
>
> I actually tested my media and I still had the problem at some point in
> my installation.
>
> Bottom line though it shouldn't just die.  Give a chance to retry,
> abort, or ignore...and continue giving that chance until the USER
> decides to abort.

I'm afraid I'll have to second this opinion.  Have seen this on 
faulty cd-rom drives which will randomly refuse to read portions
of the disc.  While no package that failed to open was critical to
the install, the fact that even one package couldn't be opened
deep-sixed the whole installation.

Possible solutions:
1) Recalculate the dependancy tree at that point, leaving out the file(s)
   that could not be read and their dependancies.

2) Just install everything as planned except for packages which appear
   to be corrupt or unreadable.  Install missing packages later.

3) Allow the person at the console to change the CD.

4) Try'n grab the missing file from an ftp site if we have the 
   network up.  For example the installer could attemt to do this
   when the failure occurs, or create a script to do this once
   the system is installed, in case of (1) or (2) above.

5) If the missing package is small enough, ask for a floppy disk
   with the package copied to it.

It seems to me that (1) or (2) in combination with (4) would be most
convenient for the person doing the install.  (3) or (5) however, would
still help and might be easier to do.

Ed

[Alan Peery]

[e-mail] wrote:
> Recalculate the dependancy tree at that point, leaving out the file(s)
>
> that could not be read and their dependancies.

I favor this option strongly.  Any of the other solutions add 
dependencies that you didn't have, and thus make things considerably 
more complicated.  Recovery from this option is also 
straightforward--install the missing bits.

It becomes tricky, however, if there is a circular dependency and one of 
the two packages is already installed correctly.  (If RPM allows 
circular dependencies at all...)  I suppose the next step then would be 
the removal of the other package and all dependent packages.

Alan

[Alan Peery]

Mike A. Harris wrote:
> I suspect the media is bad.  Please run mediacheck and test your 
> media first.  I don't see why this problem would occur only under 
> VMware personally.  Even though we don't support VMware 
> officially, problems like this still shouldn't happen.

How could media that passes an md5sum be bad?  I've not burned physical 
media, and I'm "mounting" the cdrom in a virtual drive...

Checking again, all ISO images pass /usr/lib/anaconda-runtime/checkisomd5.

The install has failed several times at exactly the same point:
1) Redhat 8.x running VMWAre 4.x, twice.  640Mb on main machine, 192 on 
virtual, almost no load on system
2) Windows XP,  running VMWare 3.x, 512Mb on laptop, 192 on visual 
machine, 20% cpu on average

Alan

[Chuck Wolber]

> (If RPM allows circular dependencies at all...)

You apparently haven't had to deal with removing yp-tools and ypbind...

-Chuck

[Mike A. Harris]

On Thu, 23 Oct 2003, Alan Peery wrote:
> > I suspect the media is bad.  Please run mediacheck and test your 
> > media first.  I don't see why this problem would occur only under 
> > VMware personally.  Even though we don't support VMware 
> > officially, problems like this still shouldn't happen.
>
> How could media that passes an md5sum be bad?

I never said media that passes an md5sum is bad.  Please point 
out where I said that.
> I've not burned physical media, and I'm "mounting" the cdrom in
> a virtual drive...

Makes no difference really.  If the MD5sums of the CDROM or ISO 
image don't match what we supply, then it is corrupt, either bad 
media, or broken download, corrupt hard disk or something.  If 
your images/CDs/whatever *DO* match our official images, then 
that isn't the problem - and that is what I was asking you for.
> Checking again, all ISO images pass /usr/lib/anaconda-runtime/checkisomd5.
>
> The install has failed several times at exactly the same point:
> Redhat 8.x running VMWAre 4.x, twice.  640Mb on main machine, 192 on
>
> virtual, almost no load on system
> 2) Windows XP,  running VMWare 3.x, 512Mb on laptop, 192 on visual 
> machine, 20% cpu on average

Well we do hope that the distribution installs inside VMware, we 
do not ever officially test installation into VMware, and it's of 
course as you said not supported.  The type of problem you're 
reporting doesn't sound to me that it would be caused by VMware 
itself as I stated before, but you never know.  Hard to say.

Can you reproduce it with a real OS install on real hardware not
using VMware?

[Alan Peery]

Mike A. Harris wrote:
> I never said media that passes an md5sum is bad.  Please point 
> out where I said that.

You didn't.  My  fractured English was meant to say that md5sums 
computed are indentical to the MD5SUMS file, so the bytes are correct in 
the files.
> If your images/CDs/whatever *DO* match our official images, then 
> that isn't the problem - and that is what I was asking you for.

They do match.
> Can you reproduce it with a real OS install on real hardware not
> using VMware?

I'll burn these same .iso files to CDROM, and then try and load with 
them after swapping hard drives.  Alas, the laptop with XP cannot be 
reinstalled at this point.

Alan

[Alan Peery]

Duncan Morison wrote:
> Never ever succesfully installed redhat on vmware via
> the CDs... this has always happenned to me.

I have successfully installed various Redhat installations directly from 
the .iso files via VMWare in the past, but it does feel indeed like this 
is a problem between VMWare and the .iso files.

I burned the exact ISO files onto physical media, and installed them on 
another machine, and the installation went fine onto other hardware.

Alan

[Stefan Hoelldampf]

Alan Peery wrote:
> > Never ever succesfully installed redhat on vmware via
> > the CDs... this has always happenned to me.
> I have successfully installed various Redhat installations directly from 
> the .iso files via VMWare in the past, but it does feel indeed like this 
> is a problem between VMWare and the .iso files.
>
> I burned the exact ISO files onto physical media, and installed them on 
> another machine, and the installation went fine onto other hardware.

Try to use the option "Legacy emulation" for your CDROM in your virtual
machine settings if the media check or installation fails.

Regards, Stefan