Mailinglist Archive: opensuse (1264 mails)

< Previous Next >
Re: [opensuse] Are many 42.2 related (not 'official') repositories often out of sync with each other?
On 2017-06-30 21:46, Knurpht - Gertjan Lettink wrote:
Op vrijdag 30 juni 2017 20:20:12 CEST schreef Carlos E. R.:
On 2017-06-30 14:47, Carl Hartung wrote:
On Fri, 30 Jun 2017 14:07:12 +0200

Knurpht - Gertjan Lettink wrote:
26 Repos? Why? This way your breaking everything that was tested
thoroughly to give a stable experience. FWIW home: and devel: repos
are meant to break stuff ( and fix it ), not for production use.

If you want the latest and greatest, Tumbleweed should be your
choice.

For managing mutliple machines, f.e. add a repo on all machines in
one command, have a look at salt(stack). This one f.e. updates a
Centos and 4 openSUSE (2 Leap, 2 TW ) machines in one go:
salt '*' pkg.upgrade

I had the same initial reaction once I'd finished enumerating them :)

But, of course, that number includes an alternate (refresh only) and

several added automatically but not enabled:
Seeing the list, it is not really excessive. You are using mostly
specific repos for things that are only there: nvidia, cinelerra,
packman... And some repos to keep some stuff more recent, like Wine.
Some of the number are disabled, so they don't really count.

To reiterate: I am not seeking help for a 'broken' system. I'm seeking
a way to automate or semi-automate the synchronization of the two
installed systems.

There is no automated way.

There is. Use Saltstack

salt '*' cmd.run 'zypper ar WHATEVER'
salt '*' cmd.run ''zypper in dog'
does the job on both machines.
openSUSE packages are available, setup is easy.

https://software.opensuse.org/package/salt

salt
A parallel remote execution system

Salt is a distributed remote execution system used to execute commands
and query data. It was developed in order to bring the best solutions
found in the world of remote execution together and make them better,
faster and more malleable. Salt accomplishes this via its ability to
handle larger loads of information, and not just dozens, but hundreds or
even thousands of individual servers, handle them quickly and through a
simple and manageable interface.


Well, that helps if you do the operations on all the computers at the
same time, but it does not help after the fact, say when a month later
you want to have a second computer with the same list of packages from
the same repos, and you do not have a list of the operations done on the
first computer.


It would be the same set of tools as needed to replicate an installation
on same or another machine. There are tools to save rpm package list and
load it again (in YaST), but they do not consider where the rpms come
from. It is what would be needed to reconstruct a machine in case of
rebuild, too.

And it is not easy to generate a list of repo/package list, because
there is no way to query they rpm database and know from which repo a
package comes. You need some code to relate vendor strings to repos instead.
Thanks for your interest & regards,

Welcome :-)




--
Cheers / Saludos,

Carlos E. R.
(from 42.2 x86_64 "Malachite" at Telcontar)

< Previous Next >
Follow Ups