diff options
author | Jan Provaznik <jprovaznik@gitlab.com> | 2019-06-26 17:42:24 +0200 |
---|---|---|
committer | Jan Provaznik <jprovaznik@gitlab.com> | 2019-07-01 09:05:31 +0200 |
commit | 52223d59515bc16c4c412a26929d2efd76759b4f (patch) | |
tree | f8646b865b0dec84cc37559551d480d1adde23a4 /config/unicorn.rb.example.development | |
parent | 66313aec761e48f4bf28461b760daa7774025fcf (diff) | |
download | gitlab-ce-puma-init-restart.tar.gz |
Run before_fork only once on bootpuma-init-restart
This aligns behavior of before_fork with Puma where before_fork
is called only on master boot, also none of our code needs to be run
before each worker fork, instead we have a use-case for deleting
Prometheus files only on boot time.
Diffstat (limited to 'config/unicorn.rb.example.development')
-rw-r--r-- | config/unicorn.rb.example.development | 15 |
1 files changed, 8 insertions, 7 deletions
diff --git a/config/unicorn.rb.example.development b/config/unicorn.rb.example.development index 1ff55701c14..9a02d5f1007 100644 --- a/config/unicorn.rb.example.development +++ b/config/unicorn.rb.example.development @@ -25,17 +25,18 @@ run_once = true before_fork do |server, worker| if run_once + # There is a difference between Puma and Unicorn: + # - Puma calls before_fork once when booting up master process + # - Unicorn runs before_fork whenever new work is spawned + # To unify this behavior we call before_fork only once (we use + # this callback for deleting Prometheus files so for our purposes + # it makes sense to align behavior with Puma) run_once = false - # The Unicorn before_exec hook only gets called on a Unicorn - # restart, but we want to clean things up (e.g. metric files) at - # first boot. - Gitlab::Cluster::LifecycleEvents.do_master_restart + # Signal application hooks that we're about to fork + Gitlab::Cluster::LifecycleEvents.do_before_fork end - # Signal application hooks that we're about to fork - Gitlab::Cluster::LifecycleEvents.do_before_fork - # The following is only recommended for memory/DB-constrained # installations. It is not needed if your system can house # twice as many worker_processes as you have configured. |