kangTwang created FLINK-21471:
---------------------------------
Summary: jobmanger, taskmanager, and job hang abnormally, and the
blobstore directory is not deleted
Key: FLINK-21471
URL: https://issues.apache.org/jira/browse/FLINK-21471
Project: Flink
Issue Type: Bug
Components: Runtime / Task
Affects Versions: 1.11.0
Reporter: kangTwang
Fix For: 1.12.0, 1.11.0
There are currently two issues as follows:
1. When flink hangs abnormally, the directory corresponding to blobstore is not
deleted, causing the disk to burst and failing to submit the flink application
2. When the flink job application crashes abnormally, the job_id corresponding
directory in the blobstore directory is not deleted, which causes the directory
to increase continuously, causing the disk to burst
Check the flink source code, the directory will be deleted only when flink is
shut down normally, it will not be deleted if it is shut down abnormally. The
current solution is to find the running blobstore directory by checking the
flink log and manually delete the useless blobstore , Is there a good solution?
? ?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
