Re: pg on Debian servers

From: Magnus Hagander <magnus(at)hagander(dot)net>
To: rob stone <floriparob(at)gmail(dot)com>
Cc: Mark Morgan Lloyd <markMLl(dot)pgsql-general(at)telemetry(dot)co(dot)uk>, "pgsql-general(at)postgresql(dot)org" <pgsql-general(at)postgresql(dot)org>
Subject: Re: pg on Debian servers
Date: 2017-11-11 13:30:07
Message-ID: CABUevEyYFdsAp=X3nmAw4SAeagud2EM0p8UneR60mDayVtW4VQ@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Sat, Nov 11, 2017 at 2:23 PM, rob stone <floriparob(at)gmail(dot)com> wrote:

>
>
> On Sat, 2017-11-11 at 13:03 +0000, Mark Morgan Lloyd wrote:
> > Apologies for something which is distro related, but I was bitten by
> > a
> > "silly mistake"- one of my own, I hasten to say- earlier.
> >
> > Several legacy programs written in Delphi ground to a halt this
> > morning,
> > which turned out to be because a Debian system had updated its copy
> > of
> > PostgreSQL and restarted the server, which broke any live
> > connections.
> >
> > At least some versions of Delphi, not to mention other IDE/RAD tools
> > with database-aware components, don't automatically try to
> > reestablish a
> > database session that's been interrupted. In any event, an
> > unexpected
> > server restart (irrespective of all investment in UPSes etc.) has
> > the
> > potential of playing havoc on a clustered system.
> >
> > Is there any way that either the package maintainer or a site
> > administrator/programmer such as myself can mark the Postgres server
> > packages as "manual upgrade only" or similar? Or since I'm almost
> > certainly not the first person to be bitten by this, is there a
> > preferred hack in mitigation?
> >
> > --
> > Mark Morgan Lloyd
> > markMLl .AT. telemetry.co .DOT. uk
> >
> > [Opinions above are the author's, not those of his employers or
> > colleagues]
> >
> >
>
> Hello Mark,
>
> Probably caused by systemd.

Systemd has nothing to do with it, it's Debian standard to restart the
services when the binaries have changed, regardless of sysvinit or systemd.

> You can disable the postgresql service and
> re-name the script in init.d. You then have to start postgres via a
> shell script.
>

The init.d script is not used with systemd.

> You can also mark packages to be on "hold" but I don't know exactly
> what happens for major version upgrades as the current version is 9 but
> when you run an upgrade via apt it will try to install version 10 which
> is no big deal as the binaries will end up in different paths, however
>

The current version is 10. The previous version was 9.6. Version 9 was more
than 5 years ago.

And the apt system will *never* try to upgrade across a major version. You
do a new install to get the new version. An upgrade operation will put you
at the latest minor release for the currently installed version.

> libpq will be updated and that may cause a restart. I run upgrades
> without any applications running so I don't know exactly what could
> happen when using unattended upgrades.
>

libpq does get upgraded, but it does not cause restarts. A restart of a
client application using libpq must be done manually by the administrator
(unless there is specific code in the client application or it's packaging
to deal with that).

--
Magnus Hagander
Me: https://www.hagander.net/ <http://www.hagander.net/>
Work: https://www.redpill-linpro.com/ <http://www.redpill-linpro.com/>

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Clodoaldo Neto 2017-11-11 14:19:35 Fedora 25 packages not signed
Previous Message Christoph Berg 2017-11-11 13:28:05 Re: pg on Debian servers