Re: Casey & Gina 2018-03-27 <282CCF96-1D9D-4283-BCDD-255DCF566A54@icloud.com>
> I wondered why --skip-systemctl-redirect wasn't documented in the man_page for pg_ctlcluster?
We just forgot... Will fix that shortly and publish it with the next
postgresql-common release.
> I discovered it trying to figure out why running pg_ctlcluster from Saltstack on Ubuntu 16.04 was hanging - adding
thatflag resolved the issue. One of the first places I checked when running into the issue was the pg_ctlcluster man
page,so I think it would be a good idea to add there...
As just discussed on IRC, don't do that. Clusters running (or daemons
in general) should be managed by systemd and cgroups, and that switch
prevents that. At first glance, it will work, but things might behave
strangely. (And after the next reboot, systemd will start the clusters
anyway, so salt shouldn't do something else on the first install.)
That said, I don't know what's wrong there. If you can reproduce it
manually, I'd be very interested.
Christoph