summary refs log tree commit homepage
diff options
context:
space:
mode:
authorEric Wong <normalperson@yhbt.net>2009-10-29 13:15:01 -0700
committerEric Wong <normalperson@yhbt.net>2009-10-29 13:15:01 -0700
commitc7750815d56c98911c1dc4db26a8778657b45a1b (patch)
tree8973299ec8f606d23b53fc94f6ca0221ff7ce177
parentfdffd90d6dc20eafe7518b7dd125f0e09df5e2f9 (diff)
The method introduced in commit
6c8a3d3c55997978bacaecc5dbbb7d03c2fee345 to avoid killing
workers after suspend/hibernate interacted badly with the change
for OpenBSD fchmod(2) compatibility introduced with the 0.93.3
release.  This interaction lead to workers with files stuck in
the zero state to never be murdered off for timeout violations.

Additionally, the method to avoid killing processes off was
never completely reliable and has been reworked even if we
entered suspend/hibernate/STOP during client processing.

This regression was discovered during continued development of the
Rainbows! test suite (which we will bring over as it becomes ready).
-rw-r--r--lib/unicorn.rb8
1 files changed, 6 insertions, 2 deletions
diff --git a/lib/unicorn.rb b/lib/unicorn.rb
index 681b7d9..7081164 100644
--- a/lib/unicorn.rb
+++ b/lib/unicorn.rb
@@ -277,6 +277,7 @@ module Unicorn
       # are trapped.  See trap_deferred
       init_self_pipe!
       respawn = true
+      last_check = Time.now
 
       QUEUE_SIGS.each { |sig| trap_deferred(sig) }
       trap(:CHLD) { |sig_nr| awaken_master }
@@ -287,7 +288,11 @@ module Unicorn
           reap_all_workers
           case SIG_QUEUE.shift
           when nil
-            murder_lazy_workers
+            # avoid murdering workers after our master process (or the
+            # machine) comes out of suspend/hibernation
+            if (last_check + timeout) >= (last_check = Time.now)
+              murder_lazy_workers
+            end
             maintain_worker_count if respawn
             master_sleep
           when :QUIT # graceful shutdown
@@ -475,7 +480,6 @@ module Unicorn
           kill_worker(:QUIT, wpid)
           next
         end
-        stat.mode == 0100000 and next
         (diff = (Time.now - stat.ctime)) <= timeout and next
         logger.error "worker=#{worker.nr} PID:#{wpid} timeout " \
                      "(#{diff}s > #{timeout}s), killing"