flink-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From rmetz...@apache.org
Subject [1/2] flink-web git commit: Add FAQ entry for task cancellation
Date Thu, 07 Jul 2016 08:30:54 GMT
Repository: flink-web
Updated Branches:
  refs/heads/asf-site 0249f87f7 -> 0f7266b97


Add FAQ entry for task cancellation


Project: http://git-wip-us.apache.org/repos/asf/flink-web/repo
Commit: http://git-wip-us.apache.org/repos/asf/flink-web/commit/31de2a2e
Tree: http://git-wip-us.apache.org/repos/asf/flink-web/tree/31de2a2e
Diff: http://git-wip-us.apache.org/repos/asf/flink-web/diff/31de2a2e

Branch: refs/heads/asf-site
Commit: 31de2a2e6d36f288b11e4351d631a39945a44100
Parents: 0249f87
Author: Robert Metzger <rmetzger@apache.org>
Authored: Tue May 17 16:52:11 2016 +0200
Committer: Robert Metzger <rmetzger@apache.org>
Committed: Thu Jul 7 10:30:07 2016 +0200

----------------------------------------------------------------------
 faq.md | 23 +++++++++++++++++++++++
 1 file changed, 23 insertions(+)
----------------------------------------------------------------------


http://git-wip-us.apache.org/repos/asf/flink-web/blob/31de2a2e/faq.md
----------------------------------------------------------------------
diff --git a/faq.md b/faq.md
index 84824c8..238e51d 100644
--- a/faq.md
+++ b/faq.md
@@ -249,6 +249,10 @@ affected JobManager or TaskManager process.
 On Windows, the TaskManager shows a table of all processes and allows you to
 destroy a process by right its entry.
 
+Both the JobManager and TaskManager services will write signals (like SIGKILL
+and SIGTERM) into their respective log files. This can be helpful for 
+debugging issues with the stopping behavior.
+
 ### I got an OutOfMemoryException. What can I do?
 
 These exceptions occur usually when the functions in the program consume a lot
@@ -275,6 +279,11 @@ entries `taskmanager.memory.fraction` or `taskmanager.memory.size`. See
the
 [Configuration Reference]({{ site.docs-snapshot }}/setup/config.html) for details. This will
leave more memory to JVM heap,
 but may cause data processing tasks to go to disk more often.
 
+Another reason for OutOfMemoryExceptions is the use of the wrong state backend.
+By default, Flink is using a heap-based state backend for operator state in
+streaming jobs. The `RocksDBStateBackend` allows state sizes larger than the
+available heap space.
+
 ### Why do the TaskManager log files become so huge?
 
 Check the logging behavior of your jobs. Emitting logging per or tuple may be
@@ -388,6 +397,20 @@ directories in his own home directory.
 Flink creates a `.flink/` directory in the users home directory
 where it stores the Flink jar and configuration file.
 
+
+### My job is not reacting to a job cancellation?
+
+Flink is canceling a job by calling the `cancel()` method on all user tasks. Ideally,
+the tasks properly react to the call and stop what they are currently doing, so that 
+all threads can shut down.
+
+If the tasks are not reacting for a certain amount of time, Flink will start interrupting
+the thread periodically.
+
+The TaskManager logs will also contain the current stack of the method where the user 
+code is blocked. 
+
+
 ## Features
 
 ### What kind of fault-tolerance does Flink provide?


Mime
View raw message