This article describes a scenario where clustered, non-SmartStore enabled indexers are out of disk space. The solutions provided here are simple options to provide some immediate relief, which may not be applicable in all environments.
Some of these can be applied to non-clustered environments as well.
These steps are not applicable to SmartStore enabled environments.
How to identify:
maxVolumeDataSizeMB is configured, but because some or all indexes are not volume managed, it doesn't restrict the disk usage for all indexes that are not using volume (e.g.homePath = $SPLUNK_DB/xxx)WARN ProcessTracker - (child_130__RollFixMetadata) IndexConfig - idx=index_name Path coldPath='/opt/path_to/index_name/colddb' (realpath '/opt/coldstore/index_name/colddb') is inside volume=cold (path='/opt/coldstore', realpath='/opt/coldstore'), but does not reference that volume. Space used by coldPath will *not* be volume-mananged. Please check indexes.conf for configuration errors.
How to fix it:
Free up disk space by allowing the indexers to properly account calculate actual usage for all indexes, and if needed, eliminate replication of non-essential indexes across the cluster.
The steps below suggest applying proper configuration changes to indexes.conf on the Cluster Manager in the appropriate location within the manager-apps directory (or master-apps depending on where apps for your indexer are stored), then making the same changes manually (or deployed via automation such as Puppet or Ansible) to the indexers' corresponding peer-apps folder.
NOTE: Most administrators are accustomed to placing configs into $SPLUNK_HOME/etc/system/local. However, for clustered indexers, settings in $SPLUNK_HOME/etc/system/local are overridden by anything found in the indexer's peer-apps directory.
index=_internal sourcetype=splunkd "does not reference that volume" | stats values(idx)
/opt/splunk/bin/splunk enable maintenance-mode
$SPLUNK_HOME/etc/manager-apps/_cluster/local/indexes.conf or deployed via a custom app)
[volume:hotwarm]
path = /mnt/fast_disk
maxVolumeDataSizeMB = 100000[volume:cold]
path = /mnt/big_disk
maxVolumeDataSizeMB = 150000
homePath and coldPath in the [default] stanza to apply the configuration to all indexes:[default]
homePath = volume:hotwarm/$_index_name/db
coldPath = volume:cold/$_index_name/colddb
Check the stanzas for each index identified from the search above, and remove any homePath or coldPath values (if they exist) in those stanzas. This will allow the values placed in [default] to apply to these stanzas. Later, once the current issue is resolved and you have confirmed data is being saved properly to the volumes for these specific stanzas, you may wish to ensure homePath and coldPath are not defined for any other indexes.
[_internal)
repFactor = 0
One at a time on each indexer (via command line)
/opt/splunk/bin/splunk stop systemctl stop Splunkd.service$SPLUNK_HOME/etc/peer-apps directory, make the same edits to indexes.conf as were applied to the Cluster Manager in the steps above./opt/splunk/bin/splunk startsystemctl start Splunkd.service/opt/splunk/bin/splunk disable maintenance-mode

Linux commands that may be useful in troubleshooting:
Show disk usage:
du --max-depth=2 -hc /opt/splunk/var/lib/splunk
Sort by disk usage of datamodel_summary folder:
du --max-depth=2 -h|grep /opt/splunk/var/lib/splunk --color datamodel_summary|sortSplunk documentation:
| Revision | Publish Date | Comments |
|---|---|---|
1.0 |
28-Sep-2026
|
Initial Release |