Let’s say you’re running a Nutanix cluster with ESXi as a hypervisor and noticing a lot of files in your datastore and different service crashing.
You’re running a VDI workload and number of files in the datastore is exceeding 30,000 and the vpxa service is crashing. This can be critical for your infrastructure as vpxa is responsible for communication between the ESXi host and the VirtualCenter Server.
Does Nutanix give you a warning before the file limit is reached?
Absolutely yes!
The NCC health check nfs_file_count_check verifies if an NFS datastore has more than 20,000 files.
Want to run the check individually?
Login from any CVM and try running the following command:
nutanix@cvm$ ncc health_checks hypervisor_checks nfs_file_count_check
Nutanix recommends that the number of files in any datastore should be less than 20,000 if vpxa instability issues occur.
Want to manually check the number of files in any datastore?
Try running the following command from an ESXi host shellfind /vmfs/volumes/<ctr name> -follow | wc -l
Want to know about the check?
Check out the KB -1514