NonStop RPM Examples
|
|
NonStop RPM Examples |
| Home | Examples | FAQs | Documents | Download | History | Support |
Below are examples of RPM analysis. RPM update #2 contains significant performance improvements. Response times are now 10-20x faster than any real-time process monitor for NonStop servers. Overhead has also been reduced 10 times. Messaging overhead was also reduced 100 times. There are also many new display options. For more information about RPM, see the RPM Documents page.
In Figure 7 below, Real-time Process Monitor (RPM) provides analysis and reporting of all Cpus and Processes in a network using the ZOOM command. In the case below ZOOM was used with a standard Windows/Linux ANSI window. Benefits to this style of interface include:
1) RPM data stream works with any Windows or Linux workstation
(no special or extra cost client or emulator is required to run RPM).
2) RPM data stream allows you to super-size displays to include 100's of lines
of filtered output via ultra-high-performance, instantaneous, sorted views
across 1000's of Cpus and tens of millions of processes.
3) ANSI or VT100 device color encoding support well defined colors,
allowing consistent display and color-encoding for everyone.4) RPM for ANSI/VT100/TTY represents the ultimate in terms of an ultra-low-overhead,
ultra-thin client. The RPM analysis data-stream is at least 10 times more efficient than
other technologies such as HTML, XML, JAVA.
In the example below, the ZOOM command provides continuous
updating, real-time reports of both Cpus and Processes.
Both short and long term utilization of Cpus and
Processes are also displayed. This is just one of more than two dozen
new displays possible with RPM update #2.
|
Cpus
Cp hh:mm ss Busy Sec QLen Disp Disk Chit Pcb PcbX Swap MLock% Mem MB
CP%ET CPU h:mm:ss ET hh:mm:ss |
|
The first part of the report above shows CPU usage. Note that \LONDON Cpu 3 is consumed at 100 %busy. The second part of the report shows Process usage. Note that there are two very busy processes in \LONDON Cpu 3 named $WEEK1 which is 50.01 %busy and $WEEK2 at 49.87 %busy. Together these two processes are entirely consuming Cpu 3 explaining why \LONDON Cpu 3 is 100% busy.
Also note above that RPM provides columns named Busy, Cpu ms, CP%ET, Cpu h:mm:ss,
and ET h:mm:ss showing total Cpu cycles used
both in the short and long term. For example, Cpu usage is reported both in
the short term ET=10.0 seconds sample, and the long term
as time since a process was started or the time since a Cpu was
reloaded. In the example above, \LONDON Cpu 3 was 100% busy during
the last 10 second period, but \LONDON Cpu 3 was only 4% busy
since Cpu 3 was reloaded some
2613:32:39 hours,
minutes, seconds ago.
Likewise
note above how in the PROCESS report, processes \LONDON $WEEK1 and $WEEK2 were
roughly 50 %busy during the past 10 seconds,
and that both processes were roughly 50 %Busy since they were started
24:49
hours:minutes ago. Meaning these two processes have consumed
100% of Cpu 3 since they were launched 24 hours ago.
| Color | Meaning |
| $WEEK1
$WEEK2 $SRV3 |
Critical highlighted
object is greater than 50% busy Warning highlighted object is greater than 10% busy Informative highlighted object greater than 1% busy |
+SET
Version...... T0877V01.AAD (01MAR10)
ALERTS....... ON
INFO......... 1
WARNING...... 10
CRITICAL..... 50
BUSYCPU...... 0
BUSYPB....... 0
ENTRIES...... 15
ID........... $RPM (SSG PID = $RPMX)
LOGGING...... OFF
LOGFILE...... $MARS.MM.LOGFILE
OBEYESCAPE... ON
RATE......... 10 seconds
SORT......... ByNode
USECS........ OFF
SSGSHELL..... $SYSTEM.SYSTEM.SSGSHELL
TERMTYPE..... VT100
In Figure 8 output below, Real-time Process Monitor
(RPM) performs analysis and reporting on
the busiest processes in a 3 node Expand network. Unlike the display above which
used VT100;
RPM can also display data for T6530 devices.
The output below shows that \CHICAGO Cpu 1 is busy due to one
process named $SRV3
at 73.38% busy.
Output below also shows \LONDON Cpu 3 is 99.66 %busy due to two processes named
$WEEK1 and
$WEEK2 .
|
RPM Figure 8 - P \*,
ByNode
Process Cpu,Pin Busy% Name RPM
T0877(01MAR10) ET=61.0 Top Pri User |
|
| Color | Meaning |
| $SRV3
$WEEK1 $WEEK2 |
Critical - object is greater than 50% busy
at 73.38% busy Warning object is greater than 10% busy at 49.83% busy Warning object is greater than 10% busy at 49.83% busy |
+SET
Version...... T0877V01.00
(01MAR10)
ALERTS....... ON
INFO......... 1
WARNING...... 10
CRITICAL..... 50
BUSYCPU...... 0
BUSYPB....... 0
DB........... $SYSTEM.RPM.DB
EDL.......... $SYSTEM.SYSTEM.ASAP2EDL
ENTRIES...... 7
ID........... $RPM (SSG PID = $RPMX)
LOGGING...... OFF
LOGFILE...... $MARS.MM.LOGFILE
OBEYESCAPE... ON
RATE......... 10
SORT......... ByNode
USECS........ OFF
SSGSHELL..... $SYSTEM.SYSTEM.SSGSHELL
TERMTYPE..... T6530
For examples of how to run the RPM Wizard, select the following link RPMWIZ Examples.
Questions or comments - Support@NonStopRPM.Com
Last modified: Nov 12, 2009