home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
SGI Performance Co-Pilot 1.3
/
SGI Performance Co-Pilot 1.3.iso
/
dist
/
dist6.4
/
pcp.idb
/
usr
/
share
/
catman
/
u_man
/
cat1
/
pmgirix.z
/
pmgirix
Wrap
Text File
|
1997-04-03
|
10KB
|
199 lines
PPPPMMMMGGGGIIIIRRRRIIIIXXXX((((1111)))) PPPPMMMMGGGGIIIIRRRRIIIIXXXX((((1111))))
NNNNAAAAMMMMEEEE
ppppmmmmggggiiiirrrriiiixxxx - a miniature IRIX performance metrics monitor
SSSSYYYYNNNNOOOOPPPPSSSSIIIISSSS
ppppmmmmggggiiiirrrriiiixxxx [----llll] [----ccccppppuuuuddddeeeellllttttaaaa _d_e_l_t_a] [----hhhh _h_o_s_t] [----tttt _d_e_l_t_a] [----VVVV] [_p_m_g_a_d_g_e_t_s
_o_p_t_i_o_n ...]
DDDDEEEESSSSCCCCRRRRIIIIPPPPTTTTIIIIOOOONNNN
ppppmmmmggggiiiirrrriiiixxxx generates a configuration file for various IRIX performance
metrics from the Performance Co-Pilot (PCP) and feeds it to ppppmmmmggggaaaaddddggggeeeettttssss(1),
which then displays them.
The ----ccccppppuuuuddddeeeellllttttaaaa option specifies an alternative update time for the CPU
gadget(s). _d_e_l_t_a should be a real number of seconds. The ----tttt option may
be used to specify an alternative update interval for the disk and
network gadgets. The load average gadget is always updated every 5
seconds. By default the CPU gadgets are updated every 0.25 seconds and
the disk and network gadgets use the default update interval for
ppppmmmmggggaaaaddddggggeeeettttssss. For hosts on slow or congested networks it is advisable to
use a larger delta for the CPU gadgets and pass a larger delta through to
ppppmmmmggggaaaaddddggggeeeettttssss using ----tttt.
The ----hhhh option specifies that ppppmmmmggggiiiirrrriiiixxxx should generate a configuration file
and run ppppmmmmggggaaaaddddggggeeeettttssss for _h_o_s_t rather than localhost.
The display is annotated with the host name in the top left-hand corner.
The ----llll option supresses this annotation.
The ----VVVV option causes ppppmmmmggggiiiirrrriiiixxxx to print the configuration file it generates
to standard output rather than running ppppmmmmggggaaaaddddggggeeeettttssss. Brave individuals may
use this to customize the configuration file.
All command line arguments except ----ccccppppuuuuddddeeeellllttttaaaa and ----VVVV are passed through to
ppppmmmmggggaaaaddddggggeeeettttssss.
Pressing the right mouse button inside the ppppmmmmggggaaaaddddggggeeeettttssss window brings up a
popup menu which allows one to terminate ppppmmmmggggaaaaddddggggeeeettttssss. Pressing the escape
key (Esc) with the mouse pointer in the window will also terminate the
application. Pressing the middle mouse button over one of the gadgets in
the window will bring up a popup menu containing a list of actions for
the gadget, if appropriate. Most gadgets with performance metrics
driving them usually have at least one action. If there is a default
action defined for a gadget, double clicking on the gadget with the left
mouse button invokes the default action.
Actions invoked from the popup menu under mouse button two will attempt
to obtain data from the host specified using the ----hhhh option. For non-PCP
commands, some extra setting up may be required. For example, the
running ttttoooopppp(1) on a remote host may involve using xxxxhhhhoooosssstttt(1) to grant that
host access to your X server and editing your .rhost file on the remote
machine to allow logins from your account on the machine running ppppmmmmggggiiiirrrriiiixxxx.
For the non-PCP commands, xxxxoooonnnn(1) is used to run text based commands (e.g.
PPPPaaaaggggeeee 1111
PPPPMMMMGGGGIIIIRRRRIIIIXXXX((((1111)))) PPPPMMMMGGGGIIIIRRRRIIIIXXXX((((1111))))
ttttoooopppp and nnnneeeettttssssttttaaaatttt). In other cases the tools may have their own
requirements (e.g. ggggrrrr____oooossssvvvviiiieeeewwww).
Gadgets that display performance metrics' values will always have an Info
action, which will create a small information window containing the names
of the metrics being displayed by the gadget along with their current
values. This can also be displayed by positioning the mouse pointer over
a gadget and pressing the space bar. Note that each metric specification
contains the host name. This can be used to determine which host is
being monitored if several ppppmmmmggggaaaaddddggggeeeettttssss windows are visible.
CPU metrics are shown as short horizontal bars (multibars) at the left
hand side of the window. The CPUs are in groups of four, with the CPU
number increasing vertically within each group and from left to right
across the groups. If there are more than 16 CPUs, a second line of CPU
groups is added. The amount of time spent in user, system, interrupt,
graphics wait and idle states for each CPU is displayed by a bar. The
Info window for a CPU bar contains the color coding and current values
for the gadget. There are various CPU-related tools on the popup menu
under mouse button two for the CPU gadgets and their label.
The load average is displayed as a bar graph. The maximum is adjusted
dynamically to the highest load average seen while ppppmmmmggggaaaaddddggggeeeettttssss has been
running. There are some load average related tools in the popup menu
under mouse button two.
The memory bars display the amount of free memory left and the number of
pages swapped out per second over the last update interval. The default
update interval may be changed with ppppmmmmggggaaaaddddggggeeeettttssss' ----tttt ( or ----ddddeeeellllttttaaaa) option.
The network traffic bars on the right hand side of the window display the
number of bytes sent (orange) and received (aqua) on each of the
machine's network interfaces. Beware, these bars are all independently
scaled according to the historical maximum number of bytes transferred
across the interface. A steady 200 bytes in on one interface may be
indistinguishable from a steady 20K bytes in on another interface. The
Info window for a network bar will display the numbers behind the gadgets
(the maximum is also shown). The ----tttt ( or ----ddddeeeellllttttaaaa) option may be used to
control the update rate for the memory bars.
The disk performance metrics are displayed as LED gadgets, one for each
disk device. Disks on the same controller are joined together by a line.
The LEDs change colour according to how many I/O operations are performed
on the disk per second. The Info window for each LED displays the color
coding for various levels of disk activity as well as the current figure.
As activity increases, the colour changes from green to yellow to orange
to red.
FFFFIIIILLLLEEEESSSS
////vvvvaaaarrrr////ppppccccpppp////ppppmmmmnnnnssss////* default PMNS specification files
PPPPaaaaggggeeee 2222
PPPPMMMMGGGGIIIIRRRRIIIIXXXX((((1111)))) PPPPMMMMGGGGIIIIRRRRIIIIXXXX((((1111))))
SSSSEEEEEEEE AAAALLLLSSSSOOOO
ppppmmmmggggaaaaddddggggeeeettttssss(1), ppppmmmmcccchhhhaaaarrrrtttt(1), mmmmppppvvvviiiissss(1), ddddkkkkvvvviiiissss(1), ggggrrrr____oooossssvvvviiiieeeewwww(1), nnnneeeettttssssttttaaaatttt(1),
ttttoooopppp(1), xxxxoooonnnn(1), xxxxhhhhoooosssstttt(1), rrrrhhhhoooossssttttssss(4).
DDDDIIIIAAAAGGGGNNNNOOOOSSSSTTTTIIIICCCCSSSS
Both ppppmmmmggggiiiirrrriiiixxxx and ppppmmmmggggaaaaddddggggeeeettttssss may produce self-explanatory error messages.
ppppmmmmggggaaaaddddggggeeeettttssss runs as an independent child process with standard output and
error directed to the terminal that ppppmmmmggggiiiirrrriiiixxxx was started from. Error
messages and warnings from mouse button two menu actions will be sent
there.
CCCCAAAAVVVVEEEEAAAATTTTSSSS
ppppmmmmggggaaaaddddggggeeeettttssss may take a long time to start if the network link to the host
being monitored is slow or congested. In such cases it is wise to use an
update interval of several seconds. Both the ----ccccppppuuuuddddeeeellllttttaaaa and ----tttt options
should be specified.
PPPPaaaaggggeeee 3333