On Tue, 20 Oct 2020 12:30:04 +0000
"Laurent Bercot" <ska-skaware_at_skarnet.org> wrote:
> s6-rc-update always errs on the side of caution, and if it cannot be
> certain that all the services have been brought up in the correct
> order according to the dependency chains, it restarts the necessary
> parts of the graph.
I think you describe very correct the problem: you can't be certain. I
understand, that you need a default for s6-rc, something, which is
automatically applied and I think your choice is not bad.
But because it could cause a production problem (think about a master
SQL server in a high traffic situation wich kills all his caches during
restart and maybe will be down for minutes), I think there should be a
way to give s6-rc some advice manually, how to handle that situation.
Something like a manual resolve function where you can manipulate the
decisions of s6-rc to walk a better route to the desired state it wants
to achieve.
You can argue that this created too much complexity for the tool, but
it would be good enough, to have a general way to manipulate the current
state of s6-rc - more support is not needed. A good interface for the
administrator could be implemented in $SOMETOOL. So you could trick
s6-rc to assume everything is fine.
If it's also too complex, I would follow your argumentation that we
have here a corner case. However: for important servers in production,
it's a clear pain point.
Best Regards
Oli
--
Automatic-Server AG •••••
Oliver Schad
Geschäftsführer
Turnerstrasse 2
9000 St. Gallen | Schweiz
www.automatic-server.com | oliver.schad_at_automatic-server.com
Tel: +41 71 511 31 11 | Mobile: +41 76 330 03 47
Received on Tue Oct 20 2020 - 12:53:50 UTC