Recently I found some schedule jobs failed to run and there are some too many links error like below:
03-27-2017 09:00:04.109 +0800 ERROR SearchScheduler - failed to mkdir /opt/splunk/var/run/splunk/dispatch/scheduler_xxxxxsplunkxxxx_RMD0000000000000_at_1490576400_9999_5XXXXXX-9C41-4D98-8XXX-XXXXXXXXX: Too many links
Anyone knows the issue is.
The error should be related to the number of folders created in dispatch directory. For Linux, there is limitation for creating number of directories, it should be about 32000 directories. Please go to the dispatch directory and count how many directories created in there. If there are more 30,000 directories, it should be the problem for creating scheduled job artifacts files in the dispatch, you will delete some of old directories first.
You can also try to google search "Linux mkdir too many links" , you can find that failed to mkdir, it should be related to this issue.