Displaying Status and Statistics. HP HyperFabric

Chapter 5

Managing HyperFabric

Displaying Status and Statistics

Displaying Status and Statistics

You can get the status of and statistics associated with many of the

HyperFabric components by using the clic_stat command, which is described below.

The clic_stat Command

The following list contains some of the information that the clic_stat command provides:

• The current fabric map, in textual format.

• The status of one or more HyperFabric adapters.

• The global configuration information for each HyperFabric adapter and switch (if the fabric contains switches). The information includes the firmware type (8-bit or 32-bit), which is used for interoperability purposes.

• The card pair information, if you configure HyperFabric in the local failover mode (applicable to applications using HMP).

The clic_stat command can also be used to enable or disable performance statistics gathering for the DLPI driver, the firmware and

HMP.

All of the statistics that can be displayed using the clic_stat command are documented in the clic_stat (1M) man page on HP-UX 11.0 and on later HP-UX releases.

If you include /opt/clic/bin in your PATH statement, you can run the command as it is shown below. Otherwise, you must include

/opt/clic/bin as part of the command name (that is,

/opt/clic/bin/clic_stat).

To use some of this command’s parameters, you must be logged in as root (see each parameter’s description below).

The syntax is as follows: clic_stat [-p perf_level] [-d display_level] [-c adapter_ID]

[-n nodename] [-?]

117

Managing HyperFabric

Displaying Status and Statistics

118

Note that the second line in the above syntax is indented for readability purposes only. When the command is typed there should not be any indentation.

The command parameters are as follows:

• -p enables/disables performance statistics gathering according to the value of perf_level, which is one of the following:

TCP Enables DLPI driver statistics when under the TCP/IP stack.

HMP Enables Hyper Messaging

Protocol (HMP) statistics gathering.

FW

RST

Enables firmware statistics gathering.

Disables/resets the DLPI driver and HMP statistics.

Note that turning on performance statistics gathering will affect the performance of HyperFabric, by increasing CPU usage and message latency (the time it takes a message to get from one point to another).

It might also decrease the throughput of data. HP recommends that you do not specify this parameter unless you are trying to troubleshoot HyperFabric problems.

You must be logged in as root to use this parameter.

• -d specifies that you want to set the level of data displayed to display_level, which is one of the following:

NET Displays fabric component statistics.

CFG

VRID

ALL

Displays the management daemon (clic_mgmtd) configuration and statistics data.

Displays the virtual route identifier information—the

VRIDs, IP addresses, switch hopcounts, and physical routes—for each HyperFabric adapter in the local node.

Displays all available status data.

Chapter 5

Managing HyperFabric

Displaying Status and Statistics

• -c displays the statistics of the adapter identified by adapter_ID.

• -n displays statistics for the node identified by nodename.

• -s specifies that you want to get the status of the switch identified by switch_ID (assigned by clic_init or SAM). Note that this parameter is meaningful only if you enabled switch management

(through the clic_init command).

To determine the switch_ID, run the clic_stat command without specifying any parameters. This displays each switch_ID known to the system. Look through the command’s output to find the ID of the switch you are interested in.

• -? displays the online help for clic_stat.

If you do not specify any of the above parameters, the current fabric map is displayed, showing the last known status of the components.

Examples of clic_stat

Some examples of using clic_stat are as follows:

• Example 1

If the local node is bently7, and you want to enable DLPI driver statistics gathering, issue this command: clic_stat -p TCP

The generated output could look like this:

============================================================================

Date: Sat Aug 5 16:08:14 2000

Node: bently7.corp2.com

----------------------------------------------------------------------------

Performance Statistics Levels - Possible perf degradation occurring:

DLPI performance statistics enabled

============================================================================

• Example 2

Chapter 5 119

Managing HyperFabric

Displaying Status and Statistics

If the local node is bently7, and you want to display the management daemon (clic_mgmtd) configuration and statistics data, issue this command: clic_stat -d CFG

The generated output could look like this:

============================================================================

Date: Sat Aug 5 16:08:12 2000

Node: bently7.corp2.com

----------------------------------------------------------------------------

============================================================================

CLIC Management Global Status/Statistics

Current Component Versions

CLIC Management process version: 1.0

CLIC Management API version:

CLIC Driver version:

Global Management Statistics

Node failures:

1.0

1.0

Nodes active:

Command session failures:

Command active sessions:

Command total sessions:

0

0

0

Management mesh session failures:

Management mesh active sessions:

0

0

1

6

Management Global Interval Timers

Fabric mapping interval (ms): 60000

Performance Statistics Levels - Possible perf degradation occurring:

DLPI performance statistics enabled

============================================================================

• Example 3 a.

HMP in the non-local failover mode

If the local node is bently7, and you want to display the statistics for the adapter clic1 (which is on bently7), issue this command: clic_stat -c clic1

120 Chapter 5

Managing HyperFabric

Displaying Status and Statistics

The generated output could look like this:

============================================================================

Date: Sat Aug 5 16:08:30 2000

Node: bently7.corp2.com

----------------------------------------------------------------------------

Adapter ID: clic1

Instance Number:

Adapter Type:

1

4X HF2 PCI

Firmware File:

Major Num:

Mgmt process driver handle:

Version:

/opt/clic/firmware/clic_fw_hf232c

238

5

1.0

H/W Path:

Primary IP address:

Uptime time:

8/4/0/0

192.0.0.9

0 days 21 hours 1 min 3 sec

Adapter State Flags Set - Multiple flags may be set:

Configured and operational

Switch mode

F/W Error Statistics

Frame buffer overflow:

Receive on disabled endpoint:

Invalid endpoint ID:

Invalid endpoint protection key:

Interleaved gathered receive:

Interleaved multi-frame bulk messages: 0

0

0

0

0

0

NQ overflow:

Send NQ overflow:

Invalid slot key for NQ credit update: 0

DLPI QOS receive buffer shortage: 0

0

0

Link congestion events:

Max send packet retry exceeded:

Link or switch failure events:

Link or switch resume events:

0

1

0

0

Bad route detected:

Bad optional data length:

Invalid message received:

CRC error:

Invalid CRC word:

Bad frame length:

Receive buffer overflow: 0

Null packets generated on link reset: 2

0

0

0

0

0

0

Chapter 5 121

Managing HyperFabric

Displaying Status and Statistics

Firmware reset notification:

Data corruption notification:

Unsupported QOS message received:

Invalid HMP VC ID:

Invalid HMP endpoint ID:

Invalid HMP endpoint protection key:

HMP message order violation:

0

0

0

0

0

0

0

Packet drops:

Transmit side congestion events:

Receive side congestion events:

Other Misc Statistics

0

1

0

Mapping message send failures: 0

============================================================================ b.

HMP in the local failover mode

If the local node is bently7, and you want to display the statistics for the adapter clic1 (which is on bently7), issue this command: clic_stat -c clic1

The generated output could look like this:

============================================================================

Date: Sat Aug 5 16:08:30 2000

Node: bently7.corp2.com

----------------------------------------------------------------------------

Adapter ID:

Card Pair: clic1 clic0

Instance Number:

Adapter Type:

Firmware File:

Major Num:

1

4X HF2 PCI

/opt/clic/firmware/clic_fw_hf232c

238

Mgmt process driver handle:

Version:

H/W Path:

Primary IP address:

Uptime time:

5

1.0

8/4/0/0

192.0.0.9

0 days 21 hours 1 min 3 sec

Adapter State Flags Set - Multiple flags may be set:

Configured and operational

Switch mode

F/W Error Statistics

122 Chapter 5

Managing HyperFabric

Displaying Status and Statistics

Frame buffer overflow:

Receive on disabled endpoint:

Invalid endpoint ID:

Invalid endpoint protection key:

Interleaved gathered receive:

Interleaved multi-frame bulk messages: 0

NQ overflow: 0

0

0

0

0

0

Send NQ overflow: 0

Invalid slot key for NQ credit update: 0

DLPI QOS receive buffer shortage:

Link congestion events:

0

0

Max send packet retry exceeded:

Link or switch failure events:

Link or switch resume events:

Bad route detected:

1

0

0

0

Bad optional data length:

Invalid message received:

CRC error:

Invalid CRC word:

Bad frame length:

Receive buffer overflow:

Null packets generated on link reset: 2

Firmware reset notification: 0

0

0

0

0

0

0

Data corruption notification:

Unsupported QOS message received:

Invalid HMP VC ID:

Invalid HMP endpoint ID:

Invalid HMP endpoint protection key:

0

0

0

0

0

HMP message order violation:

Packet drops:

Transmit side congestion events:

Receive side congestion events:

0

0

1

0

Other Misc Statistics

Mapping message send failures: 0

============================================================================

• Example 4

If the local node is bently6, and you want to display the VRIDs, IP addresses, switch hopcounts, and physical routes for each

HyperFabric adapter in bently6, issue this command: clic_stat -d VRID

The generated output if the nodes are connected in a Point-to-Point configuration could look like this:

Chapter 5 123

Managing HyperFabric

Displaying Status and Statistics

============================================================================

Date: Sat Aug 5 16:08:12 2000

Node: bently6.corp4.com

----------------------------------------------------------------------------

============================================================================

Adapter : clic0

----------------------------------------------------------------------------

VRID IP Address Switch Hopcount Route Id Physical route

----------------------------------------------------------------------------

1 192.0.0.1

0 Point to Point Connection

2

12

192.0.0.2

192.0.0.12

0

0

Point to Point Connection

Point to Point Connection

13 192.0.0.13

0 Point to Point Connection

============================================================================

Adapter : clic2

----------------------------------------------------------------------------

VRID IP Address Switch Hopcount Route Id Physical route

----------------------------------------------------------------------------

1 192.0.0.1

0 Point to Point Connection

2

12

192.0.0.2

192.0.0.12

0

0

Point to Point Connection

Point to Point Connection

13 192.0.0.13

0 Point to Point Connection

============================================================================

124 Chapter 5

Managing HyperFabric

Displaying Status and Statistics

The generated output if the nodes are connected through a switch could look like this:

============================================================================

Date: Sat Aug 5 16:08:12 2000

1

2

12

13

Node: bently6.corp4.com

----------------------------------------------------------------------------

============================================================================

Adapter : clic0

----------------------------------------------------------------------------

VRID IP Address Switch Hopcount Route Id Physical route

----------------------------------------------------------------------------

1 192.0.0.1

1 0 0x09

2

12

192.0.0.2

192.0.0.12

1

1

0

0

0x03

0x06

13 192.0.0.13

1 0 0x00

============================================================================

Adapter : clic2

----------------------------------------------------------------------------

VRID IP Address Switch Hopcount Route Id Physical route

----------------------------------------------------------------------------

192.0.0.1

192.0.0.2

192.0.0.12

192.0.0.13

1

1

1

1

0

0

0

0

0x03

0x3d

0x00

0x3a

Chapter 5 125

Managing HyperFabric

Displaying Status and Statistics

============================================================================

• Example 5

If the local node is bently7, and you want to disable all statistics gathering on bently7, issue this command: clic_stat -p RST

The generated output could look like this:

============================================================================

Date: Sat Aug 5 16:08:35 2000

Node: bently7.corp6.com

----------------------------------------------------------------------------

No performance statistics are being gathered - Reset successful

============================================================================

126 Chapter 5