10. Appendix: "No-Writeups"
This section contains references to tools which are known to exist, but which have not been fully cataloged. If anyone wishes to author an entry for one of these tools please contact: noctools- request@merit.edu. Each mention is separated by a <form-feed> for improved readability. If you intend to actually print-out this section of the catalog, then you should probably strip-out the <ff>. tuecho.c /* * Send / receive TCP or UDP echos in any of a number of bizzare ways. * * Joel P. Bion, March 1990 * Copyright (c) 1990 cisco Systems. All rights reserved. * * This "tuecho" program is distributed in the hope that it will be * useful, but WITHOUT ANY WARRANTY; without even the implied warranty * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. * * Prompts as: * Host: -- host to send echos to -- can be name or a.b.c.d -- * Enter protocol (0 = UDP, 1 = TCP) [0]: -- UDP or TCP * Size of data portion (bytes) [100]: -- bytes in data, excluding * headers -- Number of bursts [5]: -- number of bursts of packets to * send -- Packets per burst [1]: -- packets per burst, all sent AT * ONCE -- Timeout (seconds) [2]: -- how long to wait for data * Pause interval (seconds) [0]: -- Pause interval between bursts of * frames * Type of pattern (specify = 0, increment = 1) [1]: * -- if 0 specified, allow you to specify a 16bit pattern -- as four hex digits (see below). If 1, will create a -- "incrementing", cycling pattern from 0x0000 -> 0xffff -- ->. * Enter pattern (hex value) [abcd]: -- if "0" specified above */ Availability: ftp.uu.net:/networking/cisco/tuecho.c ftp.cisco.com:tuecho.c SPY An NFS monitoring/tracing tool Availability: A postscript file describing SPY is located on ftp.uu.net:/networking/ip/nfs/spy.ps.Z NFSTRACE This is the rpcspy/nfstrace package. It is described in detail in the paper "NFS Tracing by Passive Network Monitoring", which appeared in the January, 1992 USENIX conference. You'll need either a DEC machine running ULTRIX (with the packetfilter installed in the kernel) or a Sun running SunOS 4.x (with NIT). Or you'll need to do a bit of hacking. The package differs slightly from the version in the paper: - The handle->name translation facility has been removed. It's just too fragile to include in the general release. If you need it, contact me directly and I'll be happy to mail you the code. - The output format is a wee-bit different. - The IBM-RT Enet filter version is also not included, since I seem to be the only person in the world running it. RTs are really too slow for this anyway. To configure the package, edit the makefile in the obvious (to me at least) way. Note that the not all versions of SunOS NIT have working versions of the packet timestamp mechanism. Try to set the -DSTAMPS option in the makefile, and if that doesn't work, take it out. If you are actually going to use this to gather traces, I'd like to hear from you! Please send email, and share your results/traces if your organization will allow it. I maintain a mailing list of users for updates, etc. Send me mail to be added to it. Happy tracing. Matt Blaze Department of Computer Science Princeton University 35 Olden Street Princeton, NJ 08544 mab@cs.princeton.edu 609-258-3946 Availability: ftp.uu.net:/networking/ip/nfs/nfstrace.shar (or check archie) LAMER # Lame delegation notifier # Author: Bryan Beecher # Last Modified: 6/25/92 # # To make use of this software, you need to be running the # University of Michigan release of BIND 4.8.3, or any version # of named that supports the LAME_DELEGATION patches posted to # USENET. The U-M release is available via anonymous ftp from # terminator.cc.umich.edu:/unix/dns/bind4.8.3.tar.Z. # # You must also have a copy of query(1) and host(1). These # are also available via anonymous ftp in the aforementioned # place. # ------------------------------------------------------------- # ------------------------------------------------------------- # handle arguments # ------------------------------------------------------------- # -d <day> # This flag is used to append a dot-day suffix to the LOGFILE. # Handy where log files are kept around for the last week # and contain a day suffix. # # -f <logfile> # Change the LOGFILE value altogether. # # -w # Count up all of the DNS statistics for the whole week. # # -v # Be verbose. # # -t # Test mode. Do not send mail to the lame delegation # hostmasters. Availability: ftp.uu.net:/networking/ip/dns/lamer.tar.Z (or check archie) HOST host - look up host names using domain server SYNOPSIS host [-v] [-a] [-t querytype] [options] name [server] host [-v] [-a] [-t querytype] [options] -l domain [server] host [-v] [options] -H [-D] [-E] [-G] domain host [-v] [options] -C domain host [-v] [options] -A host DESCRIPTION host looks for information about Internet hosts or domains. It gets this information from a set of interconnected servers that are spread across the world. By default, it simply converts between host names and Internet addresses. However, with the -t, -a and -v options, it can be used to find all of the information about hosts or domains that is maintained by the domain nameserver. /* * Extensively modified by E. Wassenaar, Nikhef-H, <e07@nikhef.nl> * * The officially maintained source of this program is available * via anonymous ftp from machine 'ftp.nikhef.nl' [192.16.199.1] * in the directory '/pub/network' as 'host.tar.Z' * * Also available in this directory are patched versions of the * BIND 4.8.3 nameserver and resolver library which you may need * to fully exploit the features of this program, although they * are not mandatory. See the file 'README_FIRST' for details. * * You are kindly requested to report bugs and make suggestions * for improvements to the author at the given email address, * and to not re-distribute your own modifications to others. */ /* * New features * * - Major overhaul of the whole code. * - Very rigid error checking, with more verbose error messages. * - Zone listing section completely rewritten. * - It is now possible to do recursive listings into subdomains. * - Maintain resource record statistics during zone listings. * - Maintain count of hosts during zone listings. * - Exploit multiple server addresses if available. * - Option to exploit only primary server for zone transfers. * - Option to exclude info from names that do not reside in a domain. * - Implement timeout handling during connect and read. * - Write resource record output to optional logfile. * - Special MB tracing by recursively expanding MR and MG records. * - Special mode to check SOA records at each nameserver for domain. * - Special mode to check inverse mappings of host addresses. * - Code is extensively documented. */ PINGs Many many versions of the PING program exist. Each implementation has its own set of additional features. Here are a few more PINGs that are worth taking a look at. Version on ftp.cc.berkeley.edu:pub/ping: This version has duplicate packet detection, Record Route, ability to specify data pattern for packets, flood pinging, an interval option, Multicast support, etc. Version on nikhefh.nikhef.nl:/pub/network/rping.tar.Z: 'rping' is just like 'ping', but only a single probe packet is sent to test the reachability of a destination. As an option, the loose source routing facility is used to show the roundtrip route the packet has taken. Multiple addresses of remote hosts are tried until one responds. As an option, each of multiple addresses can be probed unconditionally. Contains a patch for making loose source routing work in case you have a SUN with an OMNINET ethernet controller. VRFY vrfy.tar.Z (Version 921021) 'vrfy' is a tool to verify email addresses and mailing lists. In its simplest form it takes an address "user@domain", figures out the MX hosts for "domain", and issues the SMTP command VRFY at the primary MX host (optionally all), or at "domain" itself if no MX hosts exist. Without "domain" it goes to "localhost". More complex capabilities are: recursively expanding forward files or mailing lists, and detecting mail forwarding loops. Full-blown RFC822std11(-> 2822prop) address specifications are understood. Syntax checking can be carried out either locally or remotely. Various options are provided to exploit alternative protocol suites if necessary, and to print many forms of verbose output. Obvious limitations exist, but on average it works pretty well. Needless to say you need internet (nameserver and SMTP) access. See the man page and the extensive documentation in the source for further details. Please send comments and suggestions to Eric Wassenaar <e07@nikhef.nl> If you want to receive notification of updates, please send an email with the keyword "subscribe" in the subject or the body to the address <net-dist-request@nikhef.nl> available as: nikhefh.nikhef.nl:/pub/network/vrfy.tar.Z XNETLOAD NAME xnetload - ethernet load average display for X SYNOPSIS xnetload[-toolkitoption ...] [-scale integer] [-update seconds] [-hl color] [-highlight color] [-jumpscroll pixels] [-label string] [-nolabel] host DESCRIPTION The xnetload program displays a periodically updating histo- gram of the ethernet load average for the specified host. The resulting graph is scaled as 0% to 100%, where 0% corresponds to 0mbs and 100% corresponds to 10mbs. NOTE: The specified host must be running rpc.etherd. This program has been run using X11R4 and X11R5, under the following operating systems: SUNOS 4.1.0 SUNOS 4.1.1 ULTRIX V4.2 IRIX 3.3.2 Assuming the Imake templates and Rules are in order and in the proper place on your system, these programs should compile and link straightforward by running the following sequence: xmkmf make Then, as root, issue the following: make install make install.man Then, on your host system, (or on any other system you can rlogin or rsh into) start the etherd daemon with the following (must be root): /usr/etc/rpc.etherd le0 & where le0 is the mnemonic for the primary ethernet interface. To start the xnetload program, the following command line is suggested: ./xnetload -hl red host & where "host" is the name of any reachable network node (including LOCALHOST) that is running the etherd daemon. A small xload window should appear on your local display with nine horizontal lines. The label: "Ethernet Load %" should appear in the upper left hand corner, just below any additional title bars or other decorations provided by your window manager. If the program comes up without the nine lines, or without the "Ethernet Load" label, then either your resource file is not properly installed in the appropriate app-defaults directory, or you may have picked up the wrong xnetload image. Try re-running "make install" as root, or be sure to include the "./" in front of the command name. Good Luck! The following changes have been made to this directory since R3: o Now use Athena StripChart widget. o Understands WM_DELETE_WINDOW. o 3-26-92 Modified from xload to xnetload by Roger Smith, Sterling Software at NASA-Ames Research Center, Mountain View, Calif. rsmith@proteus.arc.nasa.gov Availability: ftp proteus.arc.nasa.gov:pub/XEnetload.tar.Z (or check archie) NETTEST nettest, nettestd - Performs client and server functions for timing data throughput The nettest and nettestd commands invoke client and server programs that are used for timing data throughput of various methods of interprocess communication. For TCP and OSI con- nections, the nettest program establishes a connection with the nettestd program, and then it does count writes of size bytes, followed by count reads of size bytes. For UDP, the nettest program performs only writes; reads are not per- formed. The nettestd program, if used with UDP connections, reads the data packets and prints a message for each data packet it receives. The number and size of the reads and writes may not correlate with the number and size of the actual data packets that are transferred; it depends on the protocol that is chosen. If you append an optional k (or K) to the size, count, or bufsize value, the number specified is multiplied by 1024. This source for nettest and nettestd are provided on an "as is" basis. Cray Research does not provide any support for this code (unless you are a customer who has purchased the UNICOS operating system). We will gladly take bug reports for nettest/nettestd. Suggested fixes are prefered to just bug reports. Changes to allow nettest/nettestd to run on other architectures are also welcomed. We will try to incorporate bugfixes and update the publicly available code, but we can make no guarantees. For copyright information, see the notice in each source file. Send bug-reports/fixes to: E-mail: dab@cray.com U.S. Mail: David Borman Cray Research, Inc. 655F Lone Oak Drive Eagan, MN 55121 Notes: 1) The -b option to nettestd has not been tested... 2) The ISO code should work on a 4.4BSD system, but the gethostinfo() routine is specific to UNICOS... Availability: ftp sgi.com:/sgi/src/nettest ETHERCK etherck is a simple program that displays Sun ethernet statistics. If you have a high percents of input errors that are due to "out of buffers", then you can run the "iepatch" script to patch a kernel that uses the Intel ethernet chip ("ie"). A back of the envelope calculation shows that a .25% input error rate gives about a 10% degradation of NFS performance if 8k packets are being used. In our environment at Legato, patching the ie buffer allocation made the input error rate drop more than 2 orders of magnitude. This was after we had applied other networking fixes (e.g., using Prestoserve, going from thin wire to twisted pair) and pushed a higher load on the server. Note that both etherck and iepatch must be run by root (or you can make etherck setgid kmem). Availability: send EMAIL to: request@legato.com with a Subject line: send unsupported etherck The following is part of the 'help' file from the Legato Email Server: This message comes to you from the request server at Legato.COM, request@Legato.COM. It received a message from you asking for help. The request server is a mail-response program. That means that you mail it a request, and it mails back the response. The request server is a very dumb program. It does not have much error checking. If you don't send it the commands that it understands, it will just answer "I don't understand you". The request server has 4 commands. Each command must be the first word on a line. The request server reads your entire message before it does anything, so you can have several different commands in a single message. The request server treats the "Subject:" header line just like any other line of the message. You can use any combination of upper and lower case letters in the commands. The request server's files are organized into a series of directories and subdirectories. Each directory has an index, and each subdirectory has an index. The top-level index gives you an overview of what is in the subdirectories, and the index for each subdirectory tells you what is in it. The server has 4 commands: "help" command: The command "help" or "send help" causes the server to send you the help file. You already know this, of course, because you are reading the help file. No other commands are honored in a message that asks for help (the server figures that you had better read the help message before you do anything else). SEND a request to Legato to get the rest of the help file! NETCK netck is a shar file that contains the sources to build "netck", a network checker that uses the rstat(3R) protocol to gather and print statistics from machines on the network. netck is useful to help understand what part of what machines are potential NFS bottlenecks. To get this file, send email to the request server with the command "send unsupported netck". Availability: same as ETHERCK (send email To: request@legato.com; subject: HELP)