Using the Probe Filter
The Probe ﬁlter determines what information is collected from DMI
components, groups, and attributes. The Probe ﬁlter determines what is
stored in the SCR Repository. You can set Probe ﬁlter extension ﬂags to
collect (on) or not collect (off) speciﬁc information from the nodes, or to
collect or not collect from newly detected components, groups, or
For a description of the Probe ﬁlter, see "scrﬁlter (1M)" on page 53 in
We have set the Probe ﬁlter to maintain a balance, in usual system
environments, between capturing too much information and not
capturing enough. We recommend that you play with SCR and tune the
Probe ﬁlter until it meets your needs. Try various settings before
releasing it to the production environment, keeping in mind the
• If you capture too much information, the process may take too much
time and use a lot of system resources.
• If you capture too little information, you might miss data that is
important to managing your system.
• The Probe ﬁlter setting is a large factor in determining the number
of nodes you can manage from a server. It determines the Repository
size and the time required to capture data from each node. A full
snapshot on a mid-range server could take more than an hour and
produce 8 MBytes of data.
• SCR does not store a snapshot unless something in it has changed
since the previous snapshot. Since certain DMI values, such as
"Used Disk Space", change frequently, Probe is set by default not to
collect such attributes. If you change it to collect such information,
every snapshot might be stored in the Repository.
If you change the Probe ﬁlter settings, you can see the effect by taking a
new snapshot and running scrdiff on the old and new snapshots. The
output of scrdiff indicates missing data with a dash (-).