new version of collect (data collector)

From: Rob Urban <urban_at_NOSPAM.rto.dec.com>
Date: Thu, 22 Oct 1998 21:49:22 +0000

Hello Digital UNIX managers,

There is a new version of 'collect' available, V1.10. See below for a
list
of new features.

Collect is a data collector for operating system and process
statistics. It
is packaged with some tools that allow graphical (if somewhat simple)
analysis of *collected* data. I emphasize *collected* (the past tense)
to
reduce the risk of misunderstanding. It is NOT capable of realtime
graphical
display. It allows the relatively cheap collection of data for later
analysis. For more information, see the README files that accompany
the kits.

Lots of things have changed in this version. The most important are:
        o reading and writing compressed datafiles is possible.
Compression
          is ENABLED by default.
        o collect now does some things to make sure it gets as reliable
as
          possible statistics. These things are: increasing priority
and
          locking itself in memory. These are ON by default.

You should read the manpage carefully for more information.

The following is the complete list of changes since the last version:

August - October, 1998 (1.09 - 1.10):
        collect:
        o '-T' bug fixed (produced "can't read from kernel")
        o fixed bug when filtering, writing to binary file with only
proc
          subsystem selected (output: not allowed to write to NULL!)
        o fixed bug with '-PP<pidlist>' on playback not working
correctly
        o by default collect will lock itself in memory (can't be
swapped
          out anymore). Additionally, collect is now compiled
statically
          so that all the necessary pages get locked. (can be disabled)
        o added ability to give collect highest priority (default).
          there is a flag (-on) to disable
        o made "couldn't get UAREA for pid XX" a log instead of error
message
          (only see with -d switch)
        o added ability to play back and merge multiple input data files
        o added ability to read and write gnuzip compressed datafiles.
          Compression is by default enabled for writing. Compressed
          input files are recognized automatically.
        o added switch (-R) to set a duration for collect to run, after
which
          it will automatically stop.
        o collect can now read from stdin using "-p -" and write to
stdout
          using "-f -".

        collgui:
        o added companion ability to select multiple data files

Collect can be found in the following places:

        ftp.digital.com:/pub/DEC/collect/ (USA)
        ftp.digital.de:/pub/DEC/collect/ (Germany)
        ftp.digital.com.au/pub/unix/tools/collect (Australia)

[please be patient if you don't see the newest version on these
servers...
they are mirrors and there is therefore a delay in propagation]

There will soon be some web pages available externally. I will announce
them when they are available.

The files are:

        collect-1.10.tar.gz - the sources, gnu-zipped and tarred
(330KB)
        COLLECT110SETLD.tar - a setld-installable binary version
(2.8MB)

as usual, you will also need perl5 if you want to use 'collgui', the
graphical front end for data interpretation. perl5 is also available
in the same directory as a setld-installable kit. There are two kits

        PERL5003SETLD.tar - Perl 5.003 for Dunix V3 and above
        PERL5004SETLD.tar - Perl 5.004 for Dunix V4 and above

If you are curious about 'collgui', you can get some (somewhat older)
screenshots of it, and of the gnuplot graphics it produces. They should
be in the same directory where you found collect. They are:

        collgui-screenshot.jpg - the control application in perl5/tk
with
                                  data selected that corresponds to the
                                  following screenshot
        gnuplot-annotated.jpg - the same as above, but with annotated
curves

I'm not aware of any bugs in 'collgui'. Please keep in mind, however,
that
it is not meant to be competition for the Polycenter Performace
Whatever. The
intent was to provide _some sort_ of a graphical tool for analysis of
collected data. The interface may not be particularly intuitive. For
those
with special needs, there is always 'cfilt' (the "back end" of
'collgui')
which does the actual work of extracting data from the text output of
'collect'. It can be used to extract just about any data that one might
be interested in. There are manpages for 'collect', 'collgui', and
'cfilt'.


As usual, if you find collect useful, I'd appreciate a note.

Rob Urban (urban_at_rto.dec.NOSPAM.com)
Received on Fri Oct 23 1998 - 22:36:26 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:38 NZDT