182 Understanding the IBM TotalStorage Open Software Family
Tivoli SAN Manager collects events received from devices. Figure 4-26 shows an
example of events logged by one of the switches.
Figure 4-26 Switch events
Events about SAN components are communicated to IBM Tivoli SAN Manager
by Agents. IBM Tivoli SAN Manager logs the events in a log that looks like
Figure 4-26 above. IBM Tivoli SAN Manager then evaluates the event, and
decides if the event is (or is not) a fault.
If the event is a fault, IBM Tivoli SAN Manager indicates the change in status by
changing the color of the object in the topology map. Here is just one example,
where we simulate the loss of the switch by powering it off. Other common faults
include a Fibre Channel cable breaking, a host crash, and an unavailable storage
Chapter 4. Storage resource management 183
Figure 4-27 shows the topology display once it is refreshed. The healthy devices
are in green, and the unhealthy devices are in red. We have also circled the
devices that turned red for clarity.
Figure 4-27 Map Showing Effects of Switch Losing Power
Device #5, the switch, is obvious. The power went out, so the switch no
All the links (connections) from the switch to the devices also went red.
The Devices 1, 2, 3, and 4 went red. These devices have no Agent on them,
are known to IBM Tivoli SAN Manager only through the switch name server,
and are unavailable now that the switch is down. IBM Tivoli SAN Manager’s
map displays this accurately.
The hosts with Agents on them (CRETE, TUNGSTEN, SENEGAL, and
DIOMEDE) remain green, as they still communicate with the Server via
TCP/IP. Only their connection link to the switch turns red.
184 Understanding the IBM TotalStorage Open Software Family
When the switch is fixed, as it powers up, it will send an event back to the IBM
Tivoli SAN Manager, who will re-query the switch (running both the topology
scanner and the attribute scanner). The topology will be refreshed to reflect the
switch being back online.
Provide various reports
The primary business purpose for IBM Tivoli SAN Manager is to keep the
storage infrastructure running to support revenue-generating actives. We have
already seen how the topology display will automatically update to reflect devices
becoming available so that network administrator can quickly respond.
Another way to improve availability is to use monitoring and reports to anticipate
devices that are beginning to fail.
Not all failures can be predicted; however, many devices may fail gradually over
time, so reporting a history of problems can help anticipate this. For example, if
you see a lot of transmission errors occurring over a period of time, you might
anticipate a component failure, and schedule maintenance testing or preemptive
replacement to minimize impact on generating revenue or other critical
Reporting is provided by NetView and includes historical and real-time reporting,
on an individual device, or on a defined group of devices.
With NetView, you have a very flexible capability to build your own reports
according to your specific needs. The reports we are interested in for IBM Tivoli
SAN Manager are reports against the objects in the MIB provided by the switch
vendor. In our lab, we used an IBM 2109 16-port switch, so we used the Brocade
MIB for the Brocade Silkworm 2800.
The data elements in the MIB can report on status (device working or not
working) and performance (x frames were transmitted over this switch port in Y
With NetView, you can display historical reports based on data collected.
Figure 4-28 on page 185 shows a report of data collected over eight ports in a
two minute interval. You can set up the data collection to look for thresholds on
various MIB values, and send a trap when defined values are reached.
Chapter 4. Storage resource management 185
Figure 4-28 Graph of # Frames Transmitted over 8 ports in a 2 minute interval
Combining the MIB objects with the canned and customized reporting from
NetView provides the storage administrator with the tools needed to help keep
the SAN running all the time.
NetView can also track MIB values in real-time. Figure 4-29 on page 186 shows
real-time monitoring of traffic on switch ports. The graph shows the number of
frames transmitted from a specific port on a particular switch over a specified
time interval. You can set the polling interval to show how often the graph will