On Wed, 2014-04-23 at 03:19 +0200, Henrik Lindberg wrote: > On 2014-22-04 8:06, Thomas Hallgren wrote: > > Would a MANUAL strategy make sense? I.e. instead of rebooting the > > master, just tell it to clear the cache (perhaps per environment). > > > > - thomas > > > Circling back to this. Andy pointed out later that the best way to do > this is to get the web environment to do a graceful restart by either > sending apache or unicorn (depending on what is used) a HUP.
That should work for Apache or Nginx. But what about the new users running a master on webrick? Those are the users that we want to protect against: * stale cache and endless issues to understand why the master doesn't pick up the manifests changes * bad performance if there's no caching at all. People are prompt to make opinions (especially when something doesn't work at first). But maybe, we just don't care about webrick (I don't remember if the webrick support will abandoned or not or is it already)? > The problem with doing the manual cache invalidation is knowing which > running instance of the master to talk to, and it would either need an > IPC mechanism, or that all instances watch the same file - and then we > are back at the complex behavior we want to avoid... Well, having at most a dozen processes watching one given file shouldn't be as harmfull than having a dozen processes watching a ton of manifest files... -- Brice Figureau My Blog: http://www.masterzen.fr/ -- You received this message because you are subscribed to the Google Groups "Puppet Developers" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/puppet-dev/1398685885.26219.32.camel%40arsenic.daysofwonder.com. For more options, visit https://groups.google.com/d/optout.
