You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Anuhan Torgonshar (JIRA)" <ji...@apache.org> on 2019/07/07 14:34:00 UTC
[jira] [Updated] (YARN-9350) Change log level to WARN when
ShellCommandExecutor.execute() occurs exception
[ https://issues.apache.org/jira/browse/YARN-9350?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Anuhan Torgonshar updated YARN-9350:
------------------------------------
Issue Type: Bug (was: Improvement)
> Change log level to WARN when ShellCommandExecutor.execute() occurs exception
> ------------------------------------------------------------------------------
>
> Key: YARN-9350
> URL: https://issues.apache.org/jira/browse/YARN-9350
> Project: Hadoop YARN
> Issue Type: Bug
> Components: yarn
> Affects Versions: 3.1.0, 2.8.5
> Reporter: Anuhan Torgonshar
> Priority: Major
> Labels: easyfix
> Attachments: NodeHealthScriptRunner.java, PrivilegedOperationExecutor.java, WindowsBasedProcessTree.java
>
>
> When method ShellCommandExecutor.execute() occurs exception , it has inconsistent log level practices in different places as following :
> {code:java}
> //hadoop-2.8.5-src\hadoop-yarn-project\hadoop-yarn\hadoop-yarn-common\src\main\java\org\apache\hadoop\yarn\util\WindowsBasedProcessTree.java
> //log statement line number:69
> //log level:error
> try {
> shellExecutor.execute();
> } catch (IOException e) {
> LOG.error(StringUtils.stringifyException(e));
> } finally {
> String output = shellExecutor.getOutput();
> if (output != null &&
> output.contains("Prints to stdout a list of processes in the task")) {
> return true;
> }
> //hadoop-2.8.5-src\hadoop-common-project\hadoop-common\src\main\java\org\apache\hadoop\util\NodeHealthScriptRunner.java
> //log statement line number:116
> //log level:warn
> try {
> shexec.execute();
> } catch (ExitCodeException e) {
> ......
> } catch (Exception e) {
> LOG.warn("Caught exception : " + e.getMessage());
> if (!shexec.isTimedOut()) {
> status = HealthCheckerExitStatus.FAILED_WITH_EXCEPTION;
> } else {
> status = HealthCheckerExitStatus.TIMED_OUT;
> }
> exceptionStackTrace = StringUtils.stringifyException(e);
> } finally {
> ......
> }
> //hadoop-2.8.5-src\hadoop-yarn-project\hadoop-yarn\hadoop-yarn-server\hadoop-yarn-server-nodemanager\src\main\java\org\apache\hadoop\yarn\server\nodemanager\containermanager\linux\privileged\PrivilegedOperationExecutor.java
> //log statement line number:179
> //log level:warn
> try {
> exec.execute();
> if (LOG.isDebugEnabled()) {
> LOG.debug("command array:");
> LOG.debug(Arrays.toString(fullCommandArray));
> LOG.debug("Privileged Execution Operation Output:");
> LOG.debug(exec.getOutput());
> }
> } catch (ExitCodeException e) {
> ......
> } catch (IOException e) {
> LOG.warn("IOException executing command: ", e);
> throw new PrivilegedOperationException(e);
> }{code}
> There are 2 similar code snippets assign WARN level, when execute() occurs exception, simultaneously only 1 code snippet chooses ERROR level for same situation. Therefore, I think this one log statement is more likely to be assigned WARN level.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org