bug#59784: [version 1.4.0rc1] Retrying a failed install fails

["Ludovic Courtès" <ludo@gnu.org>]

Ludovic Courtès <ludo@gnu.org> skribis:

> I did reproduce the issue in a VM by running “ifconfig ens3 down” in a
> tty, or by killing the ‘guix substitute’ process, to cause failure of
> ‘guix system init’.  In that case the database is indeed restored, but I
> occasionally get errors like “/gnu/store/….drv: No such file or
> directory”.

The error message that’s haunting us:

  opening file `/gnu/store/….drv': No such file or directory

comes from guix-daemon.  It happens while the client is doing an
‘add-text-to-store’ RPC to add that .drv to the store.
‘LocalStore::addTextToStore’ supposedly creates the .drv file in
/gnu/store and then reads it back (‘registerValidPath’ -> ‘addValidPath’
-> ‘readDerivation’ -> ‘readFile’): this is where it gets ENOENT.

It would suggest that the database is consistent, but that somehow
writes don’t go through the overlay FS.

More investigation is needed, but we may have to live with this bug in
1.4.0.

Ludo’.