I have ES cluster running and i am using curator 4.2.6 to run the snapshots for indices in My ES cluster.
My questions are:
IF a curator a job is running and i kept flags like wait_for_completion : True and timeout: 30 Suppose my curator job takes more than 30seconds whether the job will fail?
Is it a good practice to increase the timout:60 in curator configuration file or timout_override in action file.
That depends. Several actions will automatically override the timeout_override if you have not provided a value for it. Otherwise, yes. If the client timeout is reached, Curator will log that as an error and exit, depending on the value of continue_if_exception. Typically, the aborted actions will continue in the background (snapshots, forceMerges, restores), or at least try to complete successfully.
Generally speaking, the default 30 second timeout is acceptable, since individual action timeouts can be set with timeout_override. If you have an exceptionally overtaxed or slow cluster, then a higher timeout may be desirable. If so, increasing the timeout setting is recommended.
Several actions will automatically override the timeout_override if you have not provided a value for it
[/quote]
In my client configuration file i kept timout:30 and in my snapshot action file i didn't provided value for timout_override and kept continue_if_exception:false .
Now the curator snapshot job file will run for 30s as given in client configuration file, if it didn't completed in that time then it will use the timeout_override that is 21600s right. That means totally 21630s the job will try for completion because i kept wait_for_completion : True.
If my understanding above is correct , For starting the second curator job it will be safe to start after 21630s after the first curator job.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.