Hi Folks,
We have planned to extend resources in entire cluster deployment(SH,Idxr,IdxMaster..).
These are running RHEL 6 on VMs.
Do we really require restart of splunkd after OS is updated to latest hardware changes on fly? any advises are appreciated.
Pramodh B
Splunker Jr.
We had a OS restart of our servers to read the changes
Yes. (at least for CPUs)
When Splunk starts up it assesses the host for the available memory and CPU count.
The number of CPUs present at startup dictates how many searches can be run in parallel by Splunk.
If you add more cores, Splunk will not know this, and will limit itself to the number of cores it had when it was originally started.
Memory is less likely to be limiting, as Splunk will just let the OS handle memory allocation.
The correct (actual) CPU count and installed memory figures will also be incorrectly reported to the monitoring console - whilst this is probably not going to cause any "issues" it could be confusing.
I would suggest that you should restart Splunk if you change the allocated resources.
The only restart which should be service disrupting is the SH (assuming its a single SH).
Restarts to the CM and IDX peers should not affect user searches. (If initiated correctly)
https://docs.splunk.com/Documentation/Splunk/8.0.1/Indexer/Userollingrestart