summaryrefslogtreecommitdiffstats
path: root/wallace/__init__.py
diff options
context:
space:
mode:
authorLiutauras Adomaitis <adomaitis@kolabsystems.com>2019-08-21 13:15:22 +0300
committerLiutauras Adomaitis <adomaitis@kolabsystems.com>2019-08-21 17:17:44 +0300
commit00080f3bb1e232ddb647e619ba41dca0284772de (patch)
tree35727f69cbfd7d3ab643c0b7acdf26a1cf413d75 /wallace/__init__.py
parent4c874eaeea927d75c5e99a7126a619d4ac34b756 (diff)
downloadpykolab-00080f3bb1e232ddb647e619ba41dca0284772de.tar.gz
Set default value of --max-tasks to None, which effectively makes pool workers work forever.
Summary: That should be a workaround for issue when multiprocessing spawns a process which never gets a task and that ends up with all pool processes not getting into a state to accept new task and emails start congest. Most probably that is a deadlock situation described in https://codewithoutrules.com/2018/09/04/python-multiprocessing/ . The solution is in Python3. Reviewers: vanmeeuwen Reviewed By: vanmeeuwen Subscribers: petersen, machniak Differential Revision: https://git.kolab.org/D803
Diffstat (limited to 'wallace/__init__.py')
-rw-r--r--wallace/__init__.py2
1 files changed, 1 insertions, 1 deletions
diff --git a/wallace/__init__.py b/wallace/__init__.py
index 9d529c9..21f32d4 100644
--- a/wallace/__init__.py
+++ b/wallace/__init__.py
@@ -186,7 +186,7 @@ class WallaceDaemon:
"--max-tasks",
dest = "max_tasks",
action = "store",
- default = 10,
+ default = None,
type = int,
help = _("Number of tasks per process.")
)