Closed dvizelman closed 6 years ago
I've checked the issue and it appears that issue can be easily fixed by introducing small changes in WorkerImpl.java and workerScripts/fromMultiplePriorityQueues.lua:
WorkerImpl.java , pop():
add Long.toString(System.currentTimeMillis() parameter to evalsha
workerScripts/fromMultiplePriorityQueues.lua:
local now = ARGV[1]
local firstMsg = redis.call('ZRANGEBYSCORE', queueName, '-inf', now, 'WITHSCORES', 'LIMIT' , '0' , '1')
PRs are welcome!
When workers are defined to use RESET_TO_HIGHEST_PRIORITY next queue strategy, they are suddenly stop to recognize delayed tasks and execute all the tasks as immediate tasks