- Subscribe to RSS Feed
- Mark Question as New
- Mark Question as Read
- Float this Question for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page
What are the mount options for EXT4 for HDFS
- Labels:
-
Apache Hadoop
Created ‎12-23-2016 07:24 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
In HDP DOC
It is said
The ext4 file system may have potential data loss issues with default options because of the "delayed writes" feature
But not mentioned what options should use to mount EXT4 to avoid potential data loss
Created ‎12-23-2016 06:39 PM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
You should be able to use just defaults,noatime as ext4 mount options. For the most part, data loss concerns revolve around improper shutdown of the servers (hard reboots, etc). Data loss is always a concern when a server is improperly shutdown.
The option data=ordered is the default and provides the best balance between performance and protection. If you need more data protection you can use data=journaled.
Out of the box, ext4 generally doesn't require as many tuning options as ext3 does. Most people tend to use defaults,noatime, but you can use defaults,noatime,data=journaled if you are concerned about data protection. Be aware there are likely performance impact.
Created ‎12-23-2016 06:39 PM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
You should be able to use just defaults,noatime as ext4 mount options. For the most part, data loss concerns revolve around improper shutdown of the servers (hard reboots, etc). Data loss is always a concern when a server is improperly shutdown.
The option data=ordered is the default and provides the best balance between performance and protection. If you need more data protection you can use data=journaled.
Out of the box, ext4 generally doesn't require as many tuning options as ext3 does. Most people tend to use defaults,noatime, but you can use defaults,noatime,data=journaled if you are concerned about data protection. Be aware there are likely performance impact.
Created ‎12-26-2016 06:30 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks @Michael Young
Created ‎12-23-2016 09:22 PM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi @Huahua Wei
Please refer to following for fixing data loss:
Created ‎12-24-2016 05:05 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
The statement in recommendation is related to the configuration what @Michael Young already mentioned. @ddharam is a different issue applicable to RHEL 6, good to know too.
