On 10/13/14 11:35 AM, Christoph Berg wrote:
> Yeah, though it was pretty unexpected that there was another catalog
> version bump after beta2 :-(.
Possibly, but it's not unheard of.
There could also be other issues such as WAL format changes or extension
ABI changes that could affect the upgrade
Ok, I have two versions of postgres installed and I don't know
(remember) why. As I have nothing on the 9.4 one I don't have to care
about the update. That's a bit of a relief. Sorry for the noise.
Mickaël
2014-10-14 21:11 GMT+02:00 LEDUQUE Mickaël :
> Should I have a result like this one ?
>
> V
Should I have a result like this one ?
Ver Cluster Port Status OwnerData directory Log file
9.3 main5432 down postgres /var/lib/postgresql/9.3/main
/var/log/postgresql/postgresql-9.3-main.log
9.4 main5433 down postgres /var/lib/postgresql/9.4/main
/var/log/postgresql/
Re: LEDUQUE Mickaël 2014-10-14
> > The on-disk format of the PostgreSQL 9.4 data files has changed between
> > beta2 and beta3 (and as a consequence, the catalog version number). For
> > that
> > reason, existing PostgreSQL 9.4 clusters need to be dumped using the old
> > package version
> "CB" == Christoph Berg writes:
CB> I've now put a nice NEWS message in postgresql-9.4.NEWS, and updated
CB> the preinst failure message to show a similar message. The text is:
Yes, that is a suitable message.
-JimC
--
James Cloos OpenPGP: 0x997A9F17ED7DAEA6
Christoph Berg [2014-10-14 17:35 +0200]:
> The problem is the "et cetera". I thought about putting code like that
> into the pre and postinst, but it seems to fragile to actually try.
> It's true it'll work for 90% of the standard installations, but I fear
> 90% of the code needed would have to dea
> The on-disk format of the PostgreSQL 9.4 data files has changed between
> beta2 and beta3 (and as a consequence, the catalog version number). For that
> reason, existing PostgreSQL 9.4 clusters need to be dumped using the old
> package version, and reloaded after upgrading the packages.
>
Re: James Cloos 2014-10-13
> >> This is a rather awkward position for upgraders to be in, as we
> >> essentially cannot automate this robustly.
>
> Is it really that much of a problem to do something like:
>
> su -s /bin/sh -c 'pg_dumpall -c' - postgres >/var/tmp/old_pg94.sql
> pg_dropcluste
