 Ask Solem
|
c30bebe09d
Sleep only 0.2s not 1s after empty queue (round-trip time went from 0.03s to 0.01s, without hammering CPU)
|
%!s(int64=15) %!d(string=hai) anos |
|
Ask Solem
|
4cc234159a
Pyflakes
|
%!s(int64=15) %!d(string=hai) anos |
|
Ask Solem
|
7985de1f53
More cleanups
|
%!s(int64=15) %!d(string=hai) anos |
|
Ask Solem
|
d1fb7dda7c
scheduler wake-up debug message was not working properly because repeat silencer was local to on_iteration.
|
%!s(int64=15) %!d(string=hai) anos |
|
Ask Solem
|
4041250e89
Some documentation and nitpicking.
|
%!s(int64=15) %!d(string=hai) anos |
|
Ask Solem
|
bd213cbb71
Moved CarrotListener to celery.worker.listener and ack revoked messages.
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
0197805d65
Implemented deleting (revoking of tasks) using a broadcast message to the workers + refactored apply_async / celery.messaging
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
0e68083c8e
Don't use 100% CPU when bucket queue is empty, also print the scheduler debug logs less frequently.
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
245a5dfda0
PEP8ify
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
84eb860d99
PEP8ify: Group imports
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
d1e0743a1b
Renamed bucket_queue -> ready_queue
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
4c57b60a34
Now using a scheduler to schedule tasks with ETA/countdown. A lot more efficient.
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
abbad2657b
Remove the old periodic task system
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
30722fe8d5
Cosmetic changes
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
c1c01f8bb0
Abstract away the use of multiprocessing.get_logger()
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
c5b31df850
Implement AMQP basic.qos (Quality of Service) to set the message prefetch
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
b59be258c0
Remove unused import
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
584204ced6
celeryd now survives a restart of the AMQP server! Automatically re-establish AMQP broker connection if its lost. New settings:
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
358b463111
Database access can't happen at Thread.__init__, renamed InfinityThread ->
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
022635ff85
Try to fix a race condition with PeriodicTasks.
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
55226a5b85
PeriodicWorkController now emits a debug log message for which periodic tasks
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
29e4e49c50
More debugging logging
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
8e6c9ba22c
That's 99% coverage, probably don't want to take it further
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
0ce532c68c
Set PeriodicWorkController/Mediator as daemon threads.
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
448f4656b1
PEP8ify
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
e7e48f455c
Make a common subclass for threads continously doing work.
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
e0ade80f31
PEP8ify
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
58900c42ef
Forgot to import default-backend (closes JIRA #OPAL-224)
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
e6778be0de
unlink pid file is it's stale
|
%!s(int64=16) %!d(string=hai) anos |
|
Ask Solem
|
bc1c629591
Refactor celery/worker into multiple modules
|
%!s(int64=16) %!d(string=hai) anos |
