Brocade Web Tools Administrator's Guide - Supporting Fabric OS v5.3.0 (53-1000435-01, June 2007)

Web Tools Administrator’s Guide 233
Publication Number: 53-1000435-01
General Web Tools limitations
19
Loss of Connection Occasionally, you might see the following message when you try to retrieve data from
the switch or send a request to the switch:
Switch Status Checking
The switch is not currently accessible.
The dialog title may vary, because it indicates which module is having the problem.
This is caused by the loss of HTTP connection with the switch, due to a variety of
possible problems. Web Tools will automatically try to regain the connection. While Web
Tools is trying to regain the connection, check if your Ethernet connection is still
functioning. If the problem is not with the Ethernet connection, wait for Web Tools to
recover the connection and display the following message:
You will have to resubmit your request after closing this message.
If the temporary switch connection loss is caused by switch hot code load, or other
similar operation, Switch Explorer you are currently running can be downloaded from a
different firmware version than the new one. In this case the following message
displays:
Switch connection is restored. The firmware version you are running is not in sync with
the version currently on switch. Close your browser and re-launch Webtools.
You need to close Switch Explorer and relaunch Web Tools to reopen the connection.
Out of Memory Errors If you are managing fabrics with more than 10 switches or more than 1000 ports, or if
you are using the iSCSI Gateway module extensively, you might encounter
out-of-memory errors such as the following:
java.lang.OutOfMemoryError: Java heap space
To avoid this problem, increase the default heap size in the Java Control Panel. See
“Configuring the Java plug-in” on page 4 for instructions.
Performance Monitor If the Web browser crashes or the Performance Monitor license is lost while the
Performance Monitoring window is running, some of the Performance Monitor
resources owned by Web Tools might not be cleaned up correctly.
Workaround: You might need to use the CLI to manually delete these counters. For
example, if you detect Web Tools owned resources (using perfshoweemonitor), but you
have verified that no Web users are actually using them, use the perfdeleemonitor or
perfcleareemonitor command to free the resources.
Performance Monitor For SCSI Read, Write, or Read/Write on a LUN per Port graphs, Fabric OS 4.1.0 (and
later 4.x versions) allows you to enable only two bytes or less for the LUN value mask
setting. Fabric OS 3.1 (and later 3.x versions) allows up to three bytes. Web Tools
displays an error message if you exceed this limit.
Workaround: There is no workaround.
Performance Monitor For Brocade 24000 and 48000 directors, while monitoring the performance, if one or
all the blades turn Faulty or if they are powered off or on, then the behavior of various
monitoring graphs is as follows:
The Switch Aggregate and Blade Aggregate graphs will freeze without any updates
(about the traffic).
Workaround: Close and relaunch the graphs.
The Switch Throughput Utilization, Switch Percent Utilization, and Port Snapshot Error
graphs will show the faulty/powered off slot node in the Y-Axis of the graph.
Workaround: Launch any port selection dialog and load the graphs accordingly.
Refresh option in
browsers
When a pop-up window requesting a user response is pushed into the background and
a refresh is requested, a fatal Internet Explorer error might occur.
Workaround: Restart the browser.
TABLE 14 Web Tools limitations (Continued)
Area Details