rpvstat Command

The rpvstat command man page provides reference information for the rpvstat command.

Purpose

Displays RPV client statistics.

Syntax

rpvstat -h

rpvstat [-n] [-t] [-i Interval [-c Count] [-d]] [rpvclient_name . . .]

rpvstat -N [-t] [-I Interval [-c Count] [-d]]

rpvstat -m [-n] [-t] [rpvclient_name . . .]

rpvstat -R  [-r][rpvclient_name . . .]

rpvstat -r [-R] [rpv-device(s)…]

rpvstat -A  [-t] [-i Interval [-d] [-c Count] ] [rpv-device(s)…] |

rpvstat -C  [-t] [-i Interval [-d] [-c Count] ] [rpv-device(s)…]

Description

The rpvstat command displays statistical information available from the RPV client device including:

RPV client name
Connection status
Total number of completed reads
Total number of KBs read
Total number of read errors
Total number of pending reads
Total number of pending KBs to read
Total number of completed writes
Total number of KBs written
Total number of write errors
Total number of pending writes
Total number of pending KBs to write
Statistics for asynchronous I/O
Statistics for asynchronous I/O cache

The read and write errors are displayed together. These counters indicate the number of I/O errors returned to the application.

The rpvstat command can optionally display its I/O-related statistics on a per-network basis. A network summary option of the command displays the following additional information:

Network throughput in kilobytes per second. The throughput is calculated per interval time specified by the user while in monitoring mode.
The rpvstat command can also display the highest recorded values for the pending statistics. These historical high water mark numbers are:
Maximum number of pending reads per network
Maximum number of pending kilobytes to read per network
Maximum number of pending writes per network
Maximum number of pending kilobytes to write per network

These statistics are reported on a separate display and include the additional statistic:

Number of retried I/O operations (both reads and writes). This count records the number of I/O retries that have occurred on this network or device. This can be used as an indicator for a marginal or failing network.

You can also display the statistics for asynchronous mirroring. The rpvstat command prints overall asynchronous statistics using the –A option. To display statistics per device, you need to specify the list of devices. You can display the asynchronous IO cache information using -C option.

Table 1. Flags
Flag	Description
-h	Displays command syntax and usage.
-R	Resets counters in the RPV clients (requires root privilege).
-t	Includes date and time in display.
-n	Displays statistics for individual mirroring networks.
-N	Displays summary statistics by mirroring network, including throughput rate for each network.
-i Interval	Automatically redisplays status every <Interval> seconds. The value of the <Interval> parameter must be an integer greater than zero and less than or equal to 3600. If the <Interval> parameter is not specified, then the status information is displayed once.
-c Count	Redisplays information at the indicated interval <Count> times. The value of the <Count> parameter must be an integer greater than zero and less than or equal to 999999. If the <Interval> parameter is specified, but the <Count> parameter is not, then it re-displays indefinitely.
-m	Displays historical maximum pending values (high water mark values) and accumulated retry count.
-d	Displays applicable monitored statistics as delta amounts from prior value.
-A	Display the statistics for asynchronous I/O.
-C	Display the statistics for asynchronous I/O cache.
-r	Reset counters for the asynchronous I/O cache information. You can specify the -R and -r options together to reset all counters. Requires root access.

In monitor mode (-i) if the -d option is also specified, then some statistics (completed reads, completed writes, completed kilobyte read, completed kilobyte written, and errors) are represented as delta amounts from their previously displayed values. These statistics are prefixed with a plus sign (+) on the second and succeeding displays. A delta value is not displayed under certain circumstances, such as when an error is detected in the previous iteration, or a configuration change is made between iterations.
When a list of RPV client devices is not explicitly listed on the command line, the list of all available RPV Clients is generated at command initiation. In monitor mode, this list of RPV clients to display is not refreshed on each display loop. This means any additional RPV clients added or deleted are not recognized until the command is started again.

The -i interval is the time, in seconds, between each successive gathering and display of RPV statistics in monitor mode. This interval is not a precise measure of the elapsed time between each successive updated display. The rpvstat command obtains some of the information it displays by calling system services and has no control over the amount of time these services take to complete their processing. Larger numbers of RPVs will result in the rpvstat command taking longer to gather information and will elongate the time between successive displays in monitor mode, sometimes taking much longer than the -i interval between displays.

Table 2. Operands
Field	Value
rpvclient_name	Name of one or more RPV clients for which to display information. If no RPV client names are specified, then information for all RPV clients is displayed.

The -A option will print the following statistical information for one or more asynchronous devices.

Asynchronous device name
Asynchronous status: The status will be printed as a single character.
- A - Device is fully configured for asynchronous I/O and can accept async I/Os.
- I - Asynchronous configuration is incomplete.
- U - The device is not configured with asynchronous configuration. Hence it is acting as a synchronous device. All statistics will be printed as zero.
- X - Device status can’t be retrieved. All the remaining statistics will be printed as zero.
Total number of asynchronous remote writes completed. The writes are mirrored and complete.
Total asynchronous remote writes completed in kilobyte. The writes are mirrored and complete.
Total number of asynchronous writes pending to mirror. The writes are in the cache. These writes are complete as per LVM is concerned but not yet mirrored.
Total asynchronous writes pending to mirror in kilobyte. The writes are in the cache. These writes are complete as per LVM is concerned but not yet mirrored.
Total number of writes whose response pending. These writes are in the pending queue and not yet written to cache.
Total asynchronous writes response pending in kilobyte. These writes are in the pending queue and not yet written to cache.

The -C option will print the following statistical information about the asynchronous I/O cache. The VG name is extracted from the ODM.

Volume group name
Asynchronous status: The status will be printed as a single character.
- A - Device is fully configured for asynchronous I/O and can accept asynchronous I/Os.
- I - Asynchronous configuration is incomplete.
- U - The device is not configured with asynchronous configuration. Hence it is acting as a synchronous device. All statistics will be printed as zero.
- X - Device status can’t be retrieved. All the remaining statistics will be printed as zero
Total asynchronous write operations
Maximum cache utilization in percent
Number of pending asynchronous writes waiting for the cache flush after cache hits high water mark.
Percentage of writes waiting for the cache flush after cache hits high water mark limit.
Maximum time waited after cache hits High Water Mark in seconds.
Current Free space in Cache in kilobytes.

Notes

The count of reads and writes is accumulated on a per buffer basis. This means that if an application I/O passes a vector of buffers in a single read or write call, then instead of counting that read or write as a single I/O, it is counted as the number of buffers in the vector.
The count of completed and pending I/O kilobytes is truncated. Any fractional amount of a KB is dropped in the output display.

The cx field in the display output displays the connection status. This field can be:

Table 3. cx output
Field	Description
A number	This number is the count of active network connections between the RPV Client and its RPV Server.
Y	Indicates the connection represented by the IP address is available and functioning.
N	Indicates the connection represented by the IP address is not available.
X	Indicates the required information could not be retrieved from the device driver. Reasons for this include: the device driver is not loaded, the device is not in the available state, and the device has been deleted.

Exit Status

This command returns the following exit values:

Table 4. Exist status
Field	Description
0	No errors.
>0	An error occurred.

Examples

To display statistical information for all RPV clients, enter:
```
rpvstat
```
To display statistical information for RPV client hdisk14, enter:
```
rpvstat hdisk14
```
To reset the statistical counters in RPV client hdisk23, enter:
```
rpvstat -R hdisk23
```
To display statistical information for RPV client hdisk14 and repeat the display every 30 seconds for 12 times, enter:
```
rpvstat hdisk14 -i 30 -c 12
```
To display statistical information for all RPV clients and include detailed information by mirroring network, enter:
```
rpvstat -n
```
To display statistical information for all mirroring networks, enter:
```
rpvstat -N
```
To display statistical information on maximum pending values for all RPV clients, enter:
```
rpvstat -m
```

Files

/usr/sbin/rpvstat contains the rpvstat command.