HP StorageWorks 2/16V Brocade Web Tools Administrator's Guide - Supporting Fab - Page 255

General Web Tools limitations 19 TABLE 14 Area Web Tools limitations (Continued) Details Loss of Connection Out of Memory Errors Performance Monitor Performance Monitor Performance Monitor Refresh option in browsers 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. 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. 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. 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. 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. 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. Web Tools Administrator's Guide 233 Publication Number: 53-1000435-01