You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "mahesh kumar behera (Jira)" <ji...@apache.org> on 2020/01/16 01:42:00 UTC

[jira] [Assigned] (HIVE-22733) After disable operation log property in hive, still HS2 saving the operation log

     [ https://issues.apache.org/jira/browse/HIVE-22733?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

mahesh kumar behera reassigned HIVE-22733:
------------------------------------------


> After disable operation log property in hive, still HS2 saving the operation log
> --------------------------------------------------------------------------------
>
>                 Key: HIVE-22733
>                 URL: https://issues.apache.org/jira/browse/HIVE-22733
>             Project: Hive
>          Issue Type: Bug
>            Reporter: mahesh kumar behera
>            Assignee: mahesh kumar behera
>            Priority: Major
>
> There are few issues in this area.
>  1. If logging is disabled using hive.server2.logging.operation.enabled, then operation logs for the queries should not be generated. But the registerLoggingContext method in LogUtils, registers the logging context  even if the operation log is disabled. This causes the logs to be added by logger. The registration of query context should be done only if operation logging is enabled.
> {code:java}
>  public static void registerLoggingContext(Configuration conf) {
> -    MDC.put(SESSIONID_LOG_KEY, HiveConf.getVar(conf, HiveConf.ConfVars.HIVESESSIONID));
> -    MDC.put(QUERYID_LOG_KEY, HiveConf.getVar(conf, HiveConf.ConfVars.HIVEQUERYID));
>      if (HiveConf.getBoolVar(conf, HiveConf.ConfVars.HIVE_SERVER2_LOGGING_OPERATION_ENABLED)) {
> +      MDC.put(SESSIONID_LOG_KEY, HiveConf.getVar(conf, HiveConf.ConfVars.HIVESESSIONID));
> +      MDC.put(QUERYID_LOG_KEY, HiveConf.getVar(conf, HiveConf.ConfVars.HIVEQUERYID));
>        MDC.put(OPERATIONLOG_LEVEL_KEY, HiveConf.getVar(conf, HiveConf.ConfVars.HIVE_SERVER2_LOGGING_OPERATION_LEVEL));{code}
>  
> 2. In case of failed query, we close the operations and that deletes the logging context (appender and route) from logger for that query. But if any log is added after that, the query logs are getting added and new operation log file is getting generated for the query. This looks like issue with MCD clear. MCD clear is not removing the keys from the map. If remove is used instead of clear, its working fine.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)