About this guide This administrator guide provides information about: • Setting up HP StorageWorks Fabric Watch software • Managing your SAN via HP StorageWorks Fabric Watch software Intended audience This guide is intended for: • system administrators responsible for setting up HP StorageWorks Fibre Channel Storage Area Network (SAN) switches •...
Document conventions and symbols Table 1 Document conventions Convention Element Medium blue text: Figure 1 Cross-reference links and e-mail addresses Medium blue, underlined text Web site addresses (http://www.hp.com) Bold font • Key names • Text typed into a GUI element, such as into a box •...
HP technical support Telephone numbers for worldwide technical support are listed on the HP support web site: http://www.hp.com/support/. Collect the following information before calling: • Technical support registration number (if applicable) • Product serial numbers • Product model names and numbers •...
An introduction to Fabric Watch This chapter contains the following information: • Fabric Watch overview, page 11 • Introduction to fabric health, page 12 Fabric Watch overview Fabric Watch is an optional Storage Area Network (SAN) monitoring software for B-Series HP StorageWorks switches running Fabric OS 2.2 or higher.
Following an event, Fabric Watch adds an entry to the internal port log for an individual switch and freezes the log to ensure that detail-level information is available. • RapiTrap Following an event, Fabric Watch forwards event information to a proxy switch, which then forwards the information to a server to notify you.
Table 2 Fabric Watch classes (continued) Class Description Performance Monitor Serves as a tuning tool. Performance Monitor classes group areas that track the source and destination of traffic. Use the Performance Monitor class thresholds and alarms to determine traffic load and flow and to reallocate resources appropriately.
Fabric class areas Table 4 lists Fabric Watch areas in the Fabric class and describes each area. Table 4 Fabric class sreas Area Description Domain ID Changes Monitors forcible domain ID changes. Forcible domain ID changes occur when there is a conflict of domain IDs in a single fabric and the principal switch has to assign another domain ID to a switch.
• HP StorageWorks SAN Switch 2/8V, 2/16V and 2/16N • HP StorageWorks SAN Switch 2/32 • HP StorageWorks SAN Switch 4/32 Performance monitor class areas Table 6 lists Fabric Watch areas in the Performance Monitor class and describes each area. Table 6 Performance monitor class areas Area...
Table 7 Port class areas (continued) Area Indicates Receive (RX) Performance The percentage of maximum bandwidth consumed in packet receipts. State Changes The state of the port has changed for one of the following reasons: • The port has gone offline. •...
Table 9 Security class areas (continued) Area Indicates Invalid Signatures If a switch cannot verify the signature of a packet, the switch rejects the packet and the signature becomes invalid. Invalid Timestamps If a time interval becomes too great from the time a packet is sent to the time it is received, the timestamp of the packet becomes invalid and the switch rejects it.
Table 10 SFP class areas (continued) Area Description Transmit Power The transmit power area measures the amount of outgoing laser, in µwatts. Use this to determine the condition of the SFP. If the counter often exceeds the threshold, the SFP is deteriorating. Current The current area measures the amount of supplied current to the SFP transceiver.
Triggered event behavior If you do not want notification during each sample period from the port hardware failure to the time of its triggered repair, you can define the event behavior as When an event behavior is defined as triggered, Fabric Watch sends only one event notification when the fabric meets the criteria for the event.
Figure shows an example in which each time a signal crosses the high limit, an event occurs. The blue arrows indicate the area where the event criteria is met. In this case, there is a great deal of fluctuation. Even when the monitor is set to triggered, a number of messages are sent. Figure 1 Threshold monitoring Figure...
Specifying a time base If you specify a time base value other than none seconds minute hour , or ), Fabric Watch does not use the current data value. Instead, it calculates the difference between the current data value and the data value as it existed one time base ago.
does not trigger an event even though the absolute value of the counter reaches 4, which is well above the high threshold. Figure 5 Example without an event Event settings This section describes how Fabric Watch compares a fabric element’s data value against a threshold value to determine whether or not to trigger an event.
Below event trigger The Below event trigger generates an event when a data value becomes less than the low threshold boundary. When a buffer is defined, the data value must be below the buffer value and the low threshold. Changed event trigger Use the Changed event trigger for an element that requires “rate of change”...
Port persistence In the case of port monitoring, there is an additional factor to consider. The data collected in port monitoring can vary a lot over short time periods. Therefore, the port can become a source of frequent event messages (the data can exceed the threshold range and return to a value within the threshold range). Fabric Watch uses port persistence for a port event that requires the transition of the port into a marginal status.
The trap stores event information but does not actively send alerts. Port changes do not generate SNMP traps. RAPITrap RAPITrap is a Fabric Watch alarm that actively alerts you to events. After you enable RAPITrap, Fabric Watch forwards all event information to a designated proxy switch. The host API automatically configures the proxy switch, based on firmware version.
Switch policies Switch policies are a series of rules that define specific states for the overall switch. Fabric OS interacts with Fabric Watch using these policies. Each rule defines the number of types of errors that transitions the overall switch state into a state that is not healthy. For example, you can specify a switch policy so that if a switch has two port failures, it is considered to be in a marginal state;...
Activating and accessing Fabric Watch This chapter contains the following sections: • Activating Fabric Watch, page 29 • Accessing Fabric Watch, page 29 Activating Fabric Watch Fabric Watch must be activated on each switch individually before use. Use telnet or Brocade Advanced Web Tools to activate Fabric Watch, as described next.
• ”Telnet” on page 30 • ”Advanced Web Tools” on page 30 • ”SNMP-Based enterprise managers” on page 30 • ”Configuration file” on page 32 Telnet Use a telnet session to: • Observe the current monitors on a switch with the fwShow command. •...
Page 31
• Receive alarm notification via SNMP traps. • View and configure the mail database. NOTE: The following instructions apply to the AdvantNet MIB browser. There may be some variation in the procedures when other MIB browsers are used. To configure Fabric Watch with an SNMP-based enterprise manager, begin by connecting to the switch using a MIB browser: Open a MIB browser.
Locate the Fabric Watch OID information per the following screen: Configuration file Use a configuration file to: • Upload a configuration file, make changes in a text editor, and download the file to all switches. • Upload and download the configuration file through a telnet session or with Web Tools. Uploading and downloading a configuration file to multiple switches efficiently populates your SAN with consistent Fabric Watch settings.
The fwConfigure menu contains 12 menu items. The first 11 items correspond to the classes available for configuration. Item 12, which is the default, exits the fwConfigure application. From the list displayed, enter the number corresponding to the class that you want to configure. For example, if you enter 5, the menu corresponding to the E-Port class appears.
Table 12 Element listing information - RXPerformance area menu Heading Meaning LastVal The data value of the element at the time of the last event LastState The last detected state of the element ”Fabric watch components” on page 13 for more details about classes and areas. Step 2: Configure thresholds After you’ve identified and selected the appropriate class and areas, you can configure thresholds for those classes and areas.
Page 36
Fabric Watch redraws the element table with the selected element disabled. The second row of information about the selected element does not appear any more, and the status of the element is set to disabled, as follows: Select threshold index => : (8..29) [8] 8 Index ThresholdName Port CurVal...
4. advanced configuration To customize Fabric Watch monitoring to suit to your environment, use the advanced configuration option as follows: Enter 4 at the command prompt. The system generates output similar to the following screen. The output you see varies based on the class and area you select.
Table 13 Element listing information - Advanced Configuration Menu Heading Meaning Index A numeric identifier assigned to the element ThresholdName A string identifier assigned to the element BehaviorType Frequency of alarm notifications BehaviorInt The element behavior interval, in seconds The threshold boundary section of the Advanced Configuration menu includes the threshold information for the selected area.
Page 39
Fabric Watch displays the units of measurement (Unit), time base (Time base), low threshold (Low), high threshold (High) and buffer size (BufSize) for each column. See the following screen. In this example, a value of 80% is chosen as the custom high value for RXPerformance. The default value is : change behavior type 11 : change threshold alarm level : change behavior interval...
Step 3: Configure alarms Alarms act as a signal or alert that notifies you when a threshold has been crossed. You can configure the following types of notification settings for Fabric Watch: • Triggered A triggered behavior type signals you once, after a threshold has been crossed. Triggered is the default behavior type signal for all class areas.
Page 42
Enter the total at the prompt. See the following example: : change behavior type 11 : change threshold alarm level : change behavior interval 12 : change changed alarm : change threshold boundary level 13 : change below alarm : change custom unit 14 : change above alarm : change custom time base 15 : change inBetween alarm...
Page 43
Figure shows how to select the custom settings for the threshold alarm level for the RXPerformance area. The options are either to accept the default settings or provide custom settings. : change behavior type 11 : change threshold alarm level : change behavior interval 12 : change changed alarm : change threshold boundary level...
Table 15 describes the 18 customization options displayed at the end of the Advanced Configuration menu. Table 15 Advanced configuration options Option Effect Input information change behavior type Changes the behavior type of The element index and the a single element to either required behavior type Triggered or Continuous.
Table 15 Advanced configuration options (continued) Option Effect Input information change changed alarm Changes the notification The required notification methods method for changed event occurrences for this method, but only affects the custom column. change above alarm Changes the notification The required notification methods method for above event occurrences for this method,...
A port is not considered disabled if one of the port thresholds is still enabled. To enable all the thresholds for a port, at the command prompt enter: swd77:admin> fwconfigure --enable --port 9 Configuring notifications You can be notified of an alarm condition through a notification. The tasks for configuring notifications using Fabric Watch are: •...
Configuring email notifications In environments where it is critical that you are notified about errors quickly, you might want to use email notifications. With email notifications, you can be notified of serious errors via email or a pager, so you can react quickly.
Page 48
The following confirmation message appears: Email Alert is disabled! The system returns to the main fwMailCfg menu. 3: Enable Email Alert Enter 3 in the fwMailCfg menu (shown in Figure ) to enable email alert for a specific class. The Config Show menu (shown in Figure ) appears.
Enter the email address of the person responsible for the specific class of alerts. Fabric Watch uses the default value, located between the brackets in the prompt, as the current email address for the class. A value of NONE indicates that no email address has been provided. NOTE: Email addresses must not exceed 128 characters.
Monitor Health factors Fans Fan thresholds, faulty fans. Faulty WWN card (applies to modular switches). Switch does not have a redundant CP (applies to modular switches). Blade Faulty blades (applies to modular switches). Flash Flash thresholds. Marginal Ports Port, E-Port, optical port, and copper port thresholds. Whenever these thresholds are persistently high, the port is Marginal.
asked to provide values for each FRU alarm state and alarm action. To accept the default value for each FRU (as shown in Figure ), press Return. After you have configured a FRU alarm state and alarm action, the values apply to all FRUs of that type. For example, the values specified for a slot FRU will apply to all slots in the enclosure.
Load the appropriate MIB files. First, load the Brocade common MIB file (BRCD_v5_0.mib), followed by the Brocade software MIB file (SW_v5_2.mib). If this is successful, the system displays a screen similar Figure Figure 15 Configuring Fabric Watch using SNMP Figure 2, the MIB browser has populated the left side of the screen with a MIB tree that can be navigated.
swd77:admin> agtcfgset Customizing MIB-II system variables ... At each prompt, do one of the following: o <Return> to accept current value, o enter the appropriate new value, o <Control-D> to skip the rest of configuration, or o <Control-C> to cancel any change. To correct any input mistake: <Backspace>...
Page 54
Fabric Watch displays a screen similar to the one shown in Figure Figure 18 Example OID tree Obtain the specific identifier for the element that will be modified. To get the identifier, click the swFwThresholdTable and swFwThresholdEntry directory, and run a get operation on swFwName. A list of elements appears in which each element is preceded by an identifier.
Page 55
In this example, 83.1 is numeric identifier for the element referenced as resFlash000. Traverse the fields beneath swFwClassAreaTable and swFwThresholdTable, appending the numeric identifier from the previous step to each field before performing a get or write operation. For example, to get and modify information specific to the resFlash000 element, select one of the fields and append “83.1”...
Default threshold values This appendix lists Fabric Watch default threshold values for all classes except the FRU class, which has none. The following tables list all of the default values used for the default Fabric Watch configuration settings when running Fabric OS v5.x. Environment class Table 17 provides default settings for areas in the Environment class.
NOTE: For the 4/32 SAN Switch, there is no fan default threshold because the fans are not monitored by Fabric Watch. You can use fanShow to view the 4/32 SAN Switch fan status (OK or NOT OK). However, you cannot use fwConfigure to manipulate the threshold or alarm actions against any fans. Table 17 Environment class threshold defaults Area...
Page 59
Table 17 Environment class threshold defaults (continued) Area Description Default threshold Default alarm Threshold settings settings state Power Supply Monitors power Unit: 1/0 Changed: 0 Informative supply condition (OK/FAULTY) Below: 3 Out_of_range Time Base: none Above: 3 In_range In-Between: 0 Informative The default threshold settings for all...
Fabric class Table 18 provides default settings for areas in the Fabric class. These defaults are hardware-dependent. Table 18 Fabric Class threshold defaults Area Description Default threshold Default alarm Threshold settings settings state Domain ID Monitors forcible Unit: D_ID Changed: 0 Informative Changes DOMAIN ID...
Table 21 provides default settings for areas in the End-to-End Performance Monitor class. Table 21 End-to-End performance monitor class threshold defaults Area Description Default threshold Default alarm Threshold settings settings state End-to-End Invalid Monitors the Unit: errors Changed: 0 Informative CRC Count number of CRC Time Base: none...
Page 63
Table 22 Port Class threshold defaults (continued) Area Description Default threshold Default alarm Threshold settings settings state Primitive Monitors the Unit: error(s) Changed: 0 Informative Sequence number of Time Base: minute Below: 0 Informative Protocol Error primitive Low: 0 Above: 0 Out_of_range sequence errors High: 5...
Table 23 E-Port class threshold defaults (continued) Area Description Default threshold Default alarm Threshold settings settings state Loss of Monitors the Unit: error(s) Changed: 0 Informative Synchronization number of loss of Time Base: minute Below: 0 Informative Count synchronization Low: 0 Above: 0 Out_of_range errors...
Security class Table 26 provides default settings for areas in the Security class. Table 26 Security class threshold defaults Area Description Default threshold Default alarm Threshold settings settings state API Violations Monitors API Unit: violation(s) Changed: 0 Informative violations Time Base: minute Below: 0 Informative Low: 1...
Basic Fabric Watch configuration guidelines A default Fabric Watch configuration is available for the purpose of saving setup time. As you gain familiarity with Advanced Fabric Watch features, they can be tailored to suit the fabric environment. The custom settings available in Fabric Watch provide an advanced user much needed flexibility of redefining boundary thresholds and alarm notification methods.
Using Fabric Watch with configuration files When you activate Fabric Watch, the software starts, using the default settings described in Chapter 5. You cannot alter these default settings; if the default values do not suit your specific needs, configure Fabric Watch to use more appropriate settings.
Page 72
Using Fabric Watch with configuration files...
Page 73
Index email alert environment class areas above event triggers event behavior types activating event settings with advanced web tool with telnet activating Fabric Watch fabric class areas Admin View Fabric Watch components advanced configuration FRU class areas options fsconfigure alarms fwclassinit configuring fwconfigreload...
Page 74
rack stability, warning RapiTrap related documentation resource class area security class areas setting time base to none SFP class areas SNMP capabilities SNMP trap specifying a time base Subscriber’s choice, HP switch event (error) log entry switch policies switch status policy symbols in text system requirements Table...