Archive

How to free up disk space?

Builder

I received this error this morning: "The minimum free disk space (5000MB) reached for /data/splunk/var/splunk/dispatch"

I logged into the server (Linux) found that my Splunk volume is 100% used. Unfortunately, I don't currently have the option to add additional disk capacity. So my question is, how can I free up disk space in Splunk? For example, is there a way to purge data within the main index that is over a year old or something like that? What is the recommended way to troubleshoot disk space related issues in Splunk?

Thanks!

Tags (2)
1 Solution

Builder

The default size and aging policies can definitely fill a volume. You should learn about splunk buckets, as in hot, warm and cold to understand how to handle an emergency purge like this. See this documentation:

http://docs.splunk.com/Documentation/Splunk/6.2.1/Indexer/HowSplunkstoresindexes

You can safely delete cold and warm buckets with splunk stopped, understanding that you will loose the data contained in any deleted bucket. If you set your aging/size policies appropriately, splunk should roll to frozen (which by default, actually means delete) at some point soon after this config is picked up, but I have always handled such cases by deleting cold buckets.

Some other relevant links:

http://answers.splunk.com/answers/92960/how-to-view-data-retention-settings-in-splunk.html
http://answers.splunk.com/answers/57172/how-do-i-set-retention-for-an-index-to-30-days.html

View solution in original post

Communicator

Hi Team, how about in Splunk Clustered Environment? Is it safe to do the ff. steps as mention above?
1. Splunk stop all the clustered indexers
2. Delete the files in the coldPath directory
3. Restart Splunk

0 Karma

Builder

The default size and aging policies can definitely fill a volume. You should learn about splunk buckets, as in hot, warm and cold to understand how to handle an emergency purge like this. See this documentation:

http://docs.splunk.com/Documentation/Splunk/6.2.1/Indexer/HowSplunkstoresindexes

You can safely delete cold and warm buckets with splunk stopped, understanding that you will loose the data contained in any deleted bucket. If you set your aging/size policies appropriately, splunk should roll to frozen (which by default, actually means delete) at some point soon after this config is picked up, but I have always handled such cases by deleting cold buckets.

Some other relevant links:

http://answers.splunk.com/answers/92960/how-to-view-data-retention-settings-in-splunk.html
http://answers.splunk.com/answers/57172/how-do-i-set-retention-for-an-index-to-30-days.html

View solution in original post

Builder

Thanks very much for the info! I deleted my frozen buckets and this freed up 50% of my disk space.

0 Karma

Builder

Hi echojacques,

Sure there is. You can set the frozenTimePeriodInSecs setting for the index in indexes.conf. This setting is per index, so you can choose the less important indexes. Please check the documentation for indexes.conf. If you do not specify a coldToFrozenScript , then the data will be deleted.

Please make sure that you count the seconds in a year correctly, though! Omit a zero and your data will disappear in a month, rather than a year.

You can also set maxTotalDataSizeMB, if you want.

Builder

Thanks! In addition to deleting my frozen buckets, I modified my frozenTimePeriodInSecs to a more appropriate value. I appreciate the help and the info!

0 Karma

Builder

Glad you solved your issue! Chanfoli's answer is a good one as it helps you to immediately free up some space.

0 Karma