Performance considerations
Use this information to scan a system connection without degrading performance.
Running a scan policy on an IBM Storage
Scale system
can be resource intensive and cause noticeable performance degradation on the IBM Storage
Scale system. Often, system administrators choose to
designate certain nodes or node classes for running the scans. The IBM Spectrum®
Discover interface has an input field when
creating IBM Storage
Scale connections for the
administrator to specify which nodes or node class(es) they would like to run the scan on. The value
all
will run the scan across all nodes in the cluster. Any other list (comma
separated) will be treated as a list of nodes or node classes on which to run the scan. Scan times
vary by the size of the filesystem, how many nodes are used in the scan, how many CPUs are used per
node, and whether or not the IBM Storage
Scale cluster
metadata is in flash memory.