antirez / disque

Disque is a distributed message broker
BSD 3-Clause "New" or "Revised" License
8.01k stars 538 forks source link

Received duplicated jobs #192

Open LMig2k opened 7 years ago

LMig2k commented 7 years ago

This is the steps i took.

  1. Start disque-server with one node only.
  2. Adding 10000 jobs with RETRY 300 seconds to the que_0 one by one in a single thread. When a job is added, a record(job id, create time) is inserted in the mysql db.
  3. Another process get the jobs from the que_0 one by one in a single thread. When a job is gotten, The sql is executed: update queue set executeStartTime = :now, executeStartTimeStr = CONCAT(executeStartTimeStr, :nowStr) where jobId = :jobId

Then the job will be acked, the sql executed: update queue set ack = ack + 1, endTime = :now, executeEndTimeStr = CONCAT(executeEndTimeStr, :nowStr) where jobId = :jobId The time between the job is got and acked is less than 30 seconds.

  1. After all the 10000 jobs are added and gotted, the result is: 9997 jobs with one time and 3 duplicated jobs got.

select * from queue where ack=2, got the following results:

image 5.The disque client i used is spinach