Free traces software for linux

ETrace is a configurable static port network tracing tool, similar to traceroute, but supporting ICMP, TCP, UDP and other IP protocols.

Usage:

etrace [ -BbCcnv ] [ -p profile ] [ -F config ] [ -i interface ] [ -I icmp-type ] [ -T port ] [ -U port ] [ -P protocol ] [ -r probe-count ] [ -t timeout ] [ -1 hop ] [ -h hop ] [ -m hop ] [ -A address ] [ -s port ] [ -f flags ] [ -d data ] [ -D data-file ] [ -R count ] [ -q seq] [ -w window ] target [...]

Options:

etrace has a wealth of options ranging in function from controlling output to the detailed construction of trace packets.

Profile Options:

A profile is a pre-configured list of options stored in a shared, or user specific configuration file. By defining profiles, complex etrace option sets can be easily accessed with a single command line option.

-p, --profile
Specify a profile.
-C, --clear
Clear the current list of probes. This option can be used to allow a profile to inherit options from another profile, but specify its own list of probes.
-F, --config
Specify an alternative profiles file.

Interface options

-i, --interface
Specify interface. If unspecified, etrace will examine the routing table and select the most appropriate interface for each target address.
-c, --promisc
Put in interface into promiscuous mode. As this option increases the load on the system in general, it should only be used if spoofing of source packets address is enabled with the "-A" option.
Trace Type Options
-I, --icmp
Specify an ICMP trace and the packet type to use. ICMP traces may use Echo (E or P), Timestamp (T or S), Netmask (N or M) or Info (I). The default trace probe is an ICMP Echo.
-h, --hop
Specify a specific hop to investigate.
-m, --maximum
Specify the maximum number of hops.
-r, --probes
Set the maximum number of probes to send per hop. The default is 3.
-t, --timeout
Set the maximum amount of time, in milli-seconds, to wait for a response to a probe. The default is 3000 (three seconds).

Packet Construction Options

-A, --address
Specify the source IP address of generated packets.
-s, --source
Set the source port of the generated probe packets. If unspecified, etrace uses a random high port.
-f, --flags
Specify TCP and/or IP flags. Takes a comma delimitered list of any of the following flags: RF, DF, MF, FIN, SYN, RST, PSH, ACK, URG, ECE, CWR (Default: SYN)
-d, --data
Specify the data content of generated probe packets. Standard meta-characters are recognised (e.g. "nt") as are binary values given in octal (e.g. " 00x00");
-D, --data-file
Load the data content of the generated probe packets from the specified file. Filenames beginning with @ a loaded from the etrace shared data directory (usually /usr/local/share/etrace). etrace currently ships with the following predfined packet data files: dns, ike.
-R, --random
Fill the data content of the generated probe packets with the specified number of random bytes.
-b, --badcksum
Generate and send probe packets with bad checksums.
-q, --seq
Specify the TCP sequence number.
-w, --window
Specify the TCP window size.

Output Options

-v, --verbose
Increase output verbosity.
-B, --debug
Enable debugging output.
-n, --numeric
Disable name resolution.

Examples:

etrace www.sample.com

Launches a trace ICMP Echo, the default, trace to www.sample.com. Specifiying the options "-I E" whould accomplish the same results.

etrace -T 80 www.sample.com

Similar to the previous example, except the trace is performed on TCP port 80.

etrace --udp 53 --data-file @dns ns.sample.com

Starts are trace to ns.sample.com on UDP port 53 with the trace packets containing data loaded from the file /usr/local/share/etrace/dns (a file supplied with etrace that contains a simple dns request to resolve 127.0.0.1).

etrace -p dns -p fast ns.sample.com

The default profiles shipped with etrace include "dns" (which equates to the options shown in the previous example) and "fast" (which decreases both timeouts and the number of probes sent for each hop, as well as disabling name resolution). Profiles are stackable, with latter options overriding those specified in earlier profiles.

Dnstracer determines where a given Domain Name Server (DNS) gets its information from, and follows the chain of DNS servers back to the servers which know the data.
Enhancements:
- Better handling of timeout (instead of a static value, start with five seconds and double it each retry)
- Warnings are printed if you receive a different RR type than the one you asked for (for example when you ask for an A record and receive a CNAME)
- Fixed problems when records received in the authority field are shorter than the domain the nameserver is authoritative for.
- When compiled without IPv6 support on an IPv6 capable machine the machine tried to query the IPv6 translated IPv5 address anyway.
- -C caches hosts which havent answered too.

Tie::Trace can easy print debugging with tie.

SYNOPSIS

use Tie::Trace;

my %hash;
tie %hash, "Tie::Trace";

$hash{hoge} = hogehoge; # warn Hash => Key: hoge, Value: hogehgoe at ...

my @array;
tie @aray, "Tie::Trace";
push @array, "array"; # warn Array => Point: 0, Value: array at ...

my $scalar;
tie $scalar, "Tie::Trace";
$scalar = "scalar"; # warn Scalar => Value: scalar at ...

This is usefull for print debugging. Using tie mechanism, you can see sotred value for the specified variable.

If the stored value is scalar/array/hash ref, this can check recursively.

for example;
tie %hash, "Tie::Trace";

$hash{foo} = {a => 1, b => 2}; # warn ...
$hash{foo}->{a} = 2 # warn ...

But This ignores blessed reference and tied value.

Debug::Trace is a Perl extension to trace subroutine calls.

SYNOPSIS

perl -MDebug::Trace=foo,bar yourprogram.pl

Debug::Trace instruments subroutines to provide tracing information upon every call and return.

Using Debug::Trace does not require any changes to your sources. Most often, it will be used from the command line:

perl -MDebug::Trace=foo,bar yourprogram.pl

This will have your subroutines foo() and bar() printing call and return information.

Subroutine names may be fully qualified to denote subroutines in other packages than the default main::.

By default, the trace information is output using the standard warn() function.

MODIFIERS

Modifiers can be inserted in the list of subroutines to change the default behavior of this module. All modifiers can be used in three ways:

:name to enable a specific feature.
:noname to disable a specific feature.
:name(value) to set a feature to a specific value. In general, :name is equivalent to :name(1), while :noname corresponds to :name(0).

The following modifiers are recognized:

:warn

Uses warn() to produce the trace output (default). :nowarn Sends trace output directly to STDERR.

:caller

Add basic call information to the trace message, including from where the routine was called, and by whom. This is enabled by default.

:stacktrace

Add a stack trace (call history).

:maxlen(length)

Truncate the length of the lines of trace information to length characters.

The following modifiers can be used to control the way Data::Dumper prints the values of parameters and return values. See also Data::Dumper.

:indent

Controls the style of indentation. It can be set to 0, 1, 2 or 3. Style 0 spews output without any newlines, indentation, or spaces between list items. :indent(0) is the default.

:useqq

When enabled, uses double quotes for representing string values. Whitespace other than space will be represented as [ntr], "unsafe" characters will be backslashed, and unprintable characters will be output as quoted octal integers. This is the default, use :nouseqq to disable.

:maxdepth(depth)

Can be set to a positive integer that specifies the depth beyond which which we dont print structure contents. The default is 2, which means one level of array/hashes in argument lists and return values is expanded. If you use :nomaxdepth or :maxdepth(0), nested structures are fully expanded.

:quotekeys

Controls wether hash keys are always printed quoted. The default is :noquotekeys.

sortkeys

Controls whether hash keys are dumped in sorted order. The default is :nosortkeys.

Modifiers apply only to the subroutines that follow in the list of arguments.

Devel::TraceFuncs is a Perl module with trace function calls as they happen.

SYNOPSIS

Usage:

require Devel::TraceFuncs;

max_trace_depth 5;
trace_file "foo.out";
trace_file $file_handle;

sub foo {
IN(my $f, "a message");

DEBUG "hellooo!";
}

Devel::TraceFuncs provides utilities to trace the execution of a program. It can print traces that look something like:

+-> global: 0
| +-> main::fo(4, 5) (in ./t.pm:32): now then
| | +-> main::fp(4, 5) (in ./t.pm:19)
| | | +-> main::fq() (in ./t.pm:13)
| | | | que pee doll (in ./t.pm:8)
| | | +-< main::fq() (in ./t.pm:13)
| | | cee dee (in ./t.pm:14)
| | +-< main::fp(4, 5) (in ./t.pm:19)
| | ha
| | hs (in ./t.pm:20)
| +-< main::fo(4, 5) (in ./t.pm:32): now then
| done (in ./t.pm:34)
+-< global: 0

IN

A trace begins when a function calls IN. A myd variable is passed in, such that when that function exits, the destructor for the variable is called. If this trace is to be printed, the opening line of the trace in printed at this time. Any other parameters are concatenated together, and printed on both the opening and closing lines of the trace.

I wish the syntax could be a little nicer here, but I couldnt find anything in perl that resembles Tcls uplevel or upvar commands. If I was one of the perl gods, I could have figured out a way to do something like perl5db.pl:

sub sub {
# create a new subroutine, with a myd TraceFunc object
}

DEBUG

Print some text to the trace file, at the correct depth in the trace. If the last parameter ends in "!", the arguments are printed, regardless of current depth.

trace_file

trace_file takes one argument, which is either a file name or an open file handle. All trace output will go to this file.

max_trace_depth

To avoid lots of nesting, particularly from recursive function calls, you can set the maximum depth to be traced. If this is -1 (the default), all levels of functions are traced. If it is 0, no trace output occurs, except for DEBUG statements that end in "!".

Oracle::Trace is a Perl Module for parsing Oracle Trace files.

SYNOPSIS

use Oracle::Trace;

print Oracle::Trace->new($tracefilename)->parse->test_report;

Currently the parsing and statistics are very rudimentary, and in certain matters may be fundamentally flawed - you have been warned! Expect this to improve as further development takes place.

new

Create a new object for a given Orace Trace file.
my $o_trc = Oracle::Trace->new($tracefile);

init

Initialise the object (check the tracefile).
$o_trc->init.

opentracefile

Perform basic exists/read/etc. checks on given tracefile.
Returns object or undef.
$o_trc = $o_trc->checkfile($tfile);

header

Return the Header object.
my $o_hdr = $o_trc->header;

entries

Return Entry objects which comply with given regex criteria.
my @o_ents = $o_trc->entries(type=>EXEC #d+, key=>dep, value=>0);

oids

Return the unique object ids for the currently known Entryies
my @oids = $o_trc->oids;

footer

Return the Footer object
my $o_ftr = $o_trc->footer;

test_report

Return a simple test_report of the current object.
print $o_trc->test_report(string);

mini_report

Return a simple string of descending order timings for the statements retrieved from the given objects.
my $s_str = $o_trc->mini_report($i_max, @o_objs);

Note that we use microsecond resolution for Oracle 9i and above and centisecond resolution otherwise

Devel::TraceMethods is a Perl module for tracing module calls.

SYNOPSIS

use Devel::TraceMethods qw( PackageOne PackageTwo );

Devel::TraceMethods allows you to attach a logging subroutine of your choosing to all of the methods and functions within multiple packages or classes. You can use this to trace execution. It even respects inheritance.
To enable logging, pass the name of the packages you wish to trace on the line where you use Devel::TraceMethods. It will automatically install logging for all functions in the named packages.

You can also call import() after you have use()d the module if you want to log functions and methods in another package.

You can specify per-package (or per-class) logging subroutines. For example:

Devel::TraceMethods( SomePackage => &log_one, OtherPackage => &log_two );
callback( $subroutine_reference )

By default, Devel::TraceMethods uses Carp::carp() to log a method call. You can change this with the set_logger() function. Pass a subroutine reference as the only argument, and all subsequent calls to logged methods will use the new subroutine reference instead of carp().

The first argument to the logging subroutine is the full name of the logged method. The rest of the arguments are copies of those being passed to the logged method. You can modify them in the logging subroutine without disturbing your call.