in reply to Does Net::Ping have concurrency issues?

Without looking too far into your post, I think you should probably give fping a whirl instead of rolling your own.

Makeshifts last the longest.

  • Comment on Re: Does Net::Ping have concurrency issues?

Replies are listed 'Best First'.
Re: Re: Does Net::Ping have concurrency issues?
by isotope (Deacon) on Jun 04, 2003 at 17:19 UTC
    fping is S-L-O-W... this is the best run I had: 
    $ time /usr/local/sbin/fping -r1 -g 10.100.19.1/24 -a 2> /dev/null
10.100.19.1
10.100.19.50
10.100.19.85
10.100.19.87
10.100.19.200

real    0m18.816s
user    0m0.010s
sys    0m0.000s

    [download]
    Maybe I did something wrong, but I could only get it down to one retry, and the default sure isn't zero, despite the --help claim.

    I tried Time-Slice Concurrent Ping with the modifications I posted in that thread and tuning the parameters (allow 254 outstanding pings, 1 second timeout), and this is what I get: 
    $ sudo time ./multiping.pl
Reply time for 10.100.19.1 - 3.007 seconds
Reply time for 10.100.19.50 - 3.000 seconds
Reply time for 10.100.19.85 - 2.994 seconds
Reply time for 10.100.19.87 - 2.993 seconds
Reply time for 10.100.19.200 - 2.971 seconds
PONG: 10.100.19.1
PONG: 10.100.19.50
PONG: 10.100.19.85
PONG: 10.100.19.87
PONG: 10.100.19.200
0.15user 0.02system 0:04.11elapsed 4%CPU (0avgtext+0avgdata 0maxreside
+nt)k
0inputs+0outputs (394major+191minor)pagefaults 0swaps

    [download]
    With this one, it's always less than 4.25 seconds, which is much more useful to me.

    --isotope
[reply]
[d/l]
[select]

In Section Seekers of Perl Wisdom