You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@samza.apache.org by "Lakshmi Manasa Gaduputi (Jira)" <ji...@apache.org> on 2021/12/01 01:19:00 UTC
[jira] [Created] (SAMZA-2712) AzureBlob SystemProducer: flushtimeout is not respected if main thread is uploading to azure and azure upload is stuck
Lakshmi Manasa Gaduputi created SAMZA-2712:
----------------------------------------------
Summary: AzureBlob SystemProducer: flushtimeout is not respected if main thread is uploading to azure and azure upload is stuck
Key: SAMZA-2712
URL: https://issues.apache.org/jira/browse/SAMZA-2712
Project: Samza
Issue Type: Bug
Reporter: Lakshmi Manasa Gaduputi
Assignee: Lakshmi Manasa Gaduputi
AzureBlob SystemProducer is designed such that it has a fixed number of helper/upload threads and Main thread picks up the uploading when all the upload threads are busy uploading (CallerExcutesPolicy).
However, if the main thread itself is stuck uploading to azure (very likely due to azure connection issues) then flush timeout is not respected as the main thread is stuck.
Solution: use reactor's Schedulers.boundedElastic along with subscribeOn to ensure timeout is respected for uploads and main thread responds to flush/close calls without being stuck
--
This message was sent by Atlassian Jira
(v8.20.1#820001)