Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 4 Feb 1998 19:10:52 -0800 (PST)
From:      Bill Fenner <fenner@FreeBSD.ORG>
To:        FreeBSD-gnats-submit@FreeBSD.ORG
Subject:   bin/5658: ping -R includes DNS delays in RTT
Message-ID:  <199802050310.TAA24265@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help

>Number:         5658
>Category:       bin
>Synopsis:       ping -R includes DNS delays in RTT
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          open
>Quarter:
>Keywords:
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Wed Feb  4 19:20:01 PST 1998
>Last-Modified:
>Originator:     Bill Fenner
>Organization:
Xerox
>Release:        FreeBSD 2.2.5-STABLE i386
>Environment:


>Description:

ping -R does DNS lookups on the returned path but does not take the
associated delays into account when computing RTT's for future packets.

>How-To-Repeat:

Use "ping -R" without "-n"

freefall% ping -R www.psu.edu
PING www.cac.psu.edu (146.186.157.6): 56 data bytes
64 bytes from 146.186.157.6: icmp_seq=1 ttl=245 time=575.358 ms
RR:     T1-CDROM-00-EX.US.CRL.NET (165.113.118.2)
        E0-CRL-SFO-02-E0X0.US.CRL.NET (165.113.55.2)
        pb-nap.crl.net (198.32.128.20)
        gr1-a350s1.sffca.ip.att.net (192.205.31.61)
        br1-a3120s2.sffca.ip.att.net (12.127.15.141)
        br2-h30.sffca.ip.att.net (12.127.15.150)
        br1-a350s3.wswdc.ip.att.net (12.127.15.185)
        br2-h00.wswdc.ip.att.net (12.127.15.238)
        br1-a350s2.n54ny.ip.att.net (12.127.0.6)
64 bytes from 146.186.157.6: icmp_seq=2 ttl=245 time=12848.106 ms

The DNS of all the att.net nodes is what took 12 seconds, not the
packet round trip.  This makes it look like there's something
seriously wrong with the network.

>Fix:
	
Pick one:
- Use the SO_TIMESTAMP option to make the kernel tell you when the
  packet actually arrived.  This provides correct timestamps, but doesn't
  do anything about the apparent loss rate (if you ^C and there are still
  packets in the socket buffer).
- Don't do DNS on the RR path.
- Don't send another probe until you're done printing the RR path.
>Audit-Trail:
>Unformatted:



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199802050310.TAA24265>