<link rel=stylesheet href="style.css" type="text/css">
By design, collectl gathers more data than is possible to display in an efficient, easy to read,
compact form. However, most user want their data displayed in such a form for easy
interpretation. Therefore, collectl will attempt to display all data in a single line, often
choosing a subset of the complete data for each subsystem. If the user has selected too
many systems, each line may exceed the display width and wrap. When this happens either make the
terminal window wider (maybe even using a smaller font) or choose less subsystems.
This is referred to as <i>brief</i> format and is collectl's display format of choice and
therefore the default. <i>Verbose</i> mode displays more information and results in multiple
lines of output.
Collectl will try its best to select a format consistent with the user's selection
criteria, using <i>brief</i> mode whenever possible unless explicitly told no to do so.
However there are several instances
when this mode doesn't make sense. For example, detail data will always be displayed in
<i>verbose</i> mode since it takes multiple lines for each sample. When this occurs, collectl
will automatically use <i>verbose</i> which can also be manually forced for non-detail
data using <i>--verbose</i>.
One should note that these formats are not just for interactive use and can also applied
to playback mode as well.
An additional feature of <i>brief</i> output is <i>subtotal</i> mode. If one hits the
<i>enter</i> key at
any time, the next line of output will be the subtotals (or averages on non-counters)
of all columns since
the start of collectl OR the last time the counters were zeroed. To zero the
counters enter <i>Z</i> followed by a <i>carriage return</i>.
Furthermore, if you type <i>A</i> followed by the <i>enter</i> key, the averages
will be reported. The averages/totals can also be displayed during playback
in brief mode by specifying <i>-oA</i>.
To get a better idea of what the output actually looks like, see the <a href=Examples.html>examples</a>.
You can even export your own <a href=Export.html> custom output</a>.
There are several switches that provide even more control over the look of the output in addition
to <i>--verbose</i> as described above. They are:
<li>--home moves the cursor to the home position before displaying verbose output at the start
of each interval. Only available in interactive mode, this results in a look-and-feel similar
to the <i>top</i> command.
<li>--procfilt and --slabfilt effect the output format for those respective output formats
in that these typically cause a much smaller number of processes or slabs (if used in
conjunction with --slabopts S) to be displayed, sometimes as little as a line
or two and it was felt repeating the interval header when only processes or slabs are the only data
being reporting, was too distracting and so it left off. Be sure to try it with -oT for better
<li>--top is very similar to the linux <i>top</i> command in that it shows a small subset of processes
sorted by the top consumers of the cpu, I/O or even page faults. You can even use -s to add
subsystems to the display in brief or verbose mode as well. By default this format sorts by
the top CPU users but you can choose virtually field. If you choose one of the <i>slab</i> field
names it will show the top slabs sorted by that field name.
The best way to really understand how these work in conjunction with each other is to try them
out. And don't forget you can use <i>--top</i> with playback too!
<table width=100%><tr><td align=right><i>updated Feb 21, 2011</i></td></tr></colgroup></table>