[request-tracker-maintainers] Bug#864875: debconf: frontend exits with status 10 after last upgrade when /var/cache is tmpfs

[Dominic Hargreaves]

On Wed, Feb 14, 2018 at 09:40:30PM +0200, Niko Tyni wrote:
> On Tue, Feb 13, 2018 at 11:54:33PM +0000, Dominic Hargreaves wrote:
> 
> > Apologies for the delay in responding. I recall testing this
> > at the time and not being able to reproduce - however perhaps you could
> > let us know whether changing /var/cache to not be a tmpfs makes a
> > difference? If so, I would think that would be a bug either in
> > debconf or dbconfig-common.
> 
> FWIW it looks to me like the 'set -e' command at
> 
>  https://sources.debian.org/src/request-tracker4/4.4.2-1/debian/config/#L252
> 
> is the immediate problem.
> 
> I see dbconfig-common writes out
> /etc/dbconfig-common/request-tracker4.conf when request-tracker4 first
> gets configured, with variables like dbc_install defined there. I'm
> somewhat surprised if those do not end up in the debconf database after
> calling dbc_go even if the debconf database in /var/cache got nuked in
> between, but I haven't really looked into this (and can't really promise
> to do that soon.)
> 
> So it could really be a bug in either dbconfig-common or in the config
> script or both :)

The debconf database *is* in /var/cache so I don't know how anything
can be expected to work after it's wiped. But according to the FHS
what Franz is doing is completely allowable...

CCing the Debconf developers in case they can shed any light on this.

Thanks,
Dominic.