[arch-general] makechrootpkg -d and corrupted repos

[Magnus Therning]

On Tue, Feb 15, 2011 at 12:40, Magnus Therning <magnus at therning.org> wrote:
> Over the last few weeks I've noticed that the repository created by
> 'makechrootpkg -d' regularly becomes corrupted.  Subsequent builds
> result in something like
>
>
> resolving dependencies...
> looking for inter-conflicts...
>
> Targets (2): haskell-boolean-0.0.1-1  haskell-memotrie-0.4.9-1
>
> Total Download Size:    0.00 MB
> Total Installed Size:   1.36 MB
>
> Proceed with installation? [Y/n]
> checking package integrity...
> :: File haskell-boolean-0.0.1-1-x86_64.pkg.tar.xz is corrupted. Do you
> want to delete it? [Y/n]
> :: File haskell-memotrie-0.4.9-1-x86_64.pkg.tar.xz is corrupted. Do
> you want to delete it? [Y/n]
> error: failed to commit transaction (invalid or corrupted package)
> haskell-boolean-0.0.1-1-x86_64.pkg.tar.xz is invalid or corrupted
> haskell-memotrie-0.4.9-1-x86_64.pkg.tar.xz is invalid or corrupted
> Errors occurred, no packages were upgraded.
>
>
> I then have to jump into the chroot and manually re-build the repo DB
> using repo-add.  After that I can re-issue the command and everything
> is fine.
>
> I've not done any research into this issue, but when I've asked around
> in the ArchHaskell group no one recognises this issue.  However, I
> seem to be the only one building on x86_64, but I'm nonplused why that
> would play any role.  So I thought I'd ask a little wider, has anyone
> else seen this behaviour when using 'makechrootpkg -d'?

Actually it turns out I don't need to run `repo-add` in the chroot,
it's enough to just wait a little.  So I suspect this is due to some
asynchronous behaviour in the filesystem which I bump into because I
run several `makechrootpkg` in quick succession.

/M

-- 
Magnus Therning                      OpenPGP: 0xAB4DFBA4
email: magnus at therning.org   jabber: magnus at therning.org
twitter: magthe               http://therning.org/magnus