[ARVADOS] updated: fb0f0a8eade18142e28e2691d656f262b0cdc01a

[git at public.curoverse.com]

Summary of changes:
 .../nodemanager/arvnodeman/computenode/dispatch/__init__.py | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

       via  fb0f0a8eade18142e28e2691d656f262b0cdc01a (commit)
      from  97ece5619eb80acce46ec5c7b521c08ecaaf0e86 (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.


commit fb0f0a8eade18142e28e2691d656f262b0cdc01a
Author: Tom Clegg <tom at curoverse.com>
Date:   Mon Oct 12 12:40:38 2015 -0400

    Warn about unhandled case if broken node has no ping time. refs #7286

diff --git a/services/nodemanager/arvnodeman/computenode/dispatch/__init__.py b/services/nodemanager/arvnodeman/computenode/dispatch/__init__.py
index 9ea6c32..1c828c1 100644
--- a/services/nodemanager/arvnodeman/computenode/dispatch/__init__.py
+++ b/services/nodemanager/arvnodeman/computenode/dispatch/__init__.py
@@ -325,16 +325,21 @@ class ComputeNodeMonitorActor(config.actor_class):
     def shutdown_eligible(self):
         if not self._shutdowns.window_open():
             return False
-        elif self.arvados_node is None:
+        if self.arvados_node is None:
             # Node is unpaired.
             # If it hasn't pinged Arvados after boot_fail seconds, shut it down
             return not timestamp_fresh(self.cloud_node_start_time, self.boot_fail_after)
-        elif arvados_node_missing(self.arvados_node, self.node_stale_after) and self._cloud.broken(self.cloud_node):
+        missing = arvados_node_missing(self.arvados_node, self.node_stale_after)
+        if missing and self._cloud.broken(self.cloud_node):
             # Node is paired, but Arvados says it is missing and the cloud says the node
             # is in an error state, so shut it down.
             return True
-        else:
-            return self.in_state('idle')
+        if missing is None and self._cloud.broken(self.cloud_node):
+            self._logger.warning(
+                "cloud reports broken node, but paired node %s never pinged "
+                "(bug?) -- skipped check for node_stale_after",
+                self.arvados_node['uuid'])
+        return self.in_state('idle')
 
     def consider_shutdown(self):
         next_opening = self._shutdowns.next_opening()

-----------------------------------------------------------------------


hooks/post-receive
--