Many years ago i wrote a little program called pping. its a tcp 'port
pinger' works like regular ping, except over tcp. it doesnt actually get a
file, check checks a tcp port to see if its accepting connections...
$ pping -v -h www.yahoo.com -t 3 -s 1 -c 3 -p 80
[1] www.yahoo.com:80 ... [time=6.986ms] accepting connections.
[2] www.yahoo.com:80 ... [time=6.624ms] accepting connections.
[3] www.yahoo.com:80 ... [time=7.122ms] accepting connections.
----------- www.yahoo.com:80 PPING Stats ----------
3 attempts, 3 connections, 0 failures, 0% failure rate
connect time (ms) min/avg/max 6.624/6.911/7.122
$
$ pping -h www.yahoo.com -c 1 -t 3 -p 80 -q
[1] www.yahoo.com:80 ... [time=7.807ms] accepting connections.
$ pping -h www.yahoo.com -c 1 -t 3 -p 81 -q
[1] www.yahoo.com:81 ... connection refused.
$
Its a C program that should compile on anything, except windows. :)
Should be easy to port over, but i never bothered. i actually wrote
a windows app that monitors local cpu/memory/disk usage, and a list
of remote sites, and if any of those sites fail a check, it sends mail to
someone...
If anyone wants either one ill throw it up on a web page somewere..
-mike
On Thu, 26 Feb 1998, Gasparini, Edy wrote:
> On Thursday, 26 February 1998 2:00, Michael
> Sorbera[SMTP:msorber @
ibm .
net] wrote:
> >
> > Does anyone know of a program that will monitor a web server (no ping,
> > but an actual access of the URL), and if the access doesn't work, page
> > me...
> >
>
> this is real quick and real dirty - but it works ...
>
> I wrote a simple shell script that simply telnets to a port. Let's
> assume we monitoring a web server, running on standard port 80. I telnet
> to port 80 and capture std error and output to a file. close the telnet
> session and then grep the outfile for "Connected to" which is the
> standard string you get when telnet connects to a remote system. If the
> grep is successful, then no problems - if grep fails then you know the
> telnet connection failed and the web server is in trouble - quick do
> something! Our paging service provides an email interface where you can
> send a mesg to an email address which is then passed to the pager. Bingo
> - automatic alerting. This job is then scheduled to run regularly via
> cron.
>
> two main problems with this setup: one is you have to rely on your email
> gateway to be up and running in order to receive the alert. the other is
> that the host running the script could go down (in which case you
> probably have bigger problems) but you need to have that host monitored
> via some other mechanism.
>
> > I would prefer a DOS, Win 3.X or WIN95 solution. But will go to NT or
> > UNIX if need be.
> >
>
> If you want continuous monitoring, then I would suggest using a reliable
> and stable Unix host (please, no flame wars - if you've got Bourne on
> NT, then use that if you must).
>
> Here's a copy of my script:
>
> -----start script----
> #!/bin/sh
>
> #
>
> # This script will check the status of given port of a given
> host, ie. whether
>
> # port 80 (http) is listening on the web server. It
> achieves this by simply
>
> # telnetting to the host and port and looking
> for the connected string that
>
> # telnet returns when it is successful. As
> an example:
>
> #
>
> # $ telnet hostname 80
>
> # Trying...
>
> # Connected to
> hostname.
>
> # Escape character is '^]'.
>
> #
>
> # We specifically look for the
> "Connected to" string. If it's not found then we
>
> # assume something is
> broken and then raise an alert by sending email (either
>
> # to a real
> email address or a pager via email gateway).
>
> #
>
> # We can check the status
> of multiple hosts and ports with this one script. The
>
> # variable
> controlling this is DESTHOST which is simply space delimited
>
> #
> host:port pairs. For example, to test port 80 of the web server you
> would
>
> # define DESTHOST as www:80.
>
> #
>
> # The sendmesg function is where we
> actually deliver an alert. The case
>
> # statement MUST have an entry for
> each host:port defined in DESTHOST. This way,
>
> # we can notify different
> ways for each individual
> host:port.
>
> #
>
> ###########################################################
> #################
>
> #
>
> # add each host:port to check in
> DESTHOST
>
> #
>
> DESTHOST="hostname1:port hostname2:port ..."
>
> #
>
> # no need to
> modify any of this stuff - have a look in sendmesg function!
>
> # (just
> make sure the paths are OK)
> #
>
> OUTFILE=/tmp/www.out.$$
>
> SERVER_STATUS=0 # 0=alive,
> 1=dead
>
> TELNET=/usr/ucb/telnet
>
> LOGGER=/usr/ucb/logger
>
> MAIL=/usr/bin/mailx
>
>
> MESG=""
>
> ###############################################################
> ##########
>
> # use this function to send a message when something fails.
> Ensure you define
>
> # MESG and TO correctly and that each entry in
> DESTHOST has a case here.
> # You can have multiple entries in TO (so you can send the same mesg to
> multiple
> # recipients in one hit). Remember that email aliases are defined in
> # /usr/lib/aliases, so you can either have the fully qualified email
> address
> # in here or an alias defined in aliases.
>
> #
>
> #
>
> sendmesg()
>
> {
>
> case
> ${server} in
>
> hostname1:port)
>
> SUBJ="URGENT!
> hostname1 is down! "
>
> MESG="insert some extra info here if you
> want"
>
> TO="support pager"
>
>
>
> echo ${MESG} | ${MAIL}
> -s "${SUBJ}" ${TO}
>
> ${LOGGER} -t server_check "${SUBJ}"
> #place an entry in syslog
>
> ;;
>
>
> hostname2:port)
> SUBJ="URGENT! hostname2 is down! "
> MESG="insert some extra info here if you want"
> TO="support pager someone-else"
>
> echo ${MESG} | ${MAIL} -s "${SUBJ}" ${TO}
> ${LOGGER} -t server_check "${SUBJ}" #place an entry
> in syslog
> ;;
>
> esac
>
> }
>
>
>
> #
>
> ##############################################################
> ###########
>
> # for each host:port pair listed in $DESTHOST, try and
> connect to see
>
> # if something is listening.
>
> #
>
>
>
> for server in
> ${DESTHOST}
>
> do
>
> SERVER=`echo ${server}|awk -F: '{print $1}'`
>
>
> PORT=`echo ${server}|awk -F: '{print $2}'`
>
> rm -rf
> ${OUTFILE}_${SERVER}
>
>
>
> ${TELNET} ${SERVER} ${PORT} >
> ${OUTFILE}_${SERVER} 2>1 <<EOF
>
> EOF
>
>
>
> grep "Connected to ${SERVER}"
> ${OUTFILE}_${SERVER} >/dev/null
>
> status=$? #quick, grab the
> status of that last command
>
>
>
> if [ ${status} != 0 ] #which means the
> connect failed
>
> then
>
> SERVER_STATUS=1
>
>
> sendmesg
>
> else
>
> echo "$server OK"
>
>
> ${LOGGER} -t server_check "$server responds OK"
>
> fi
>
>
>
> rm -rf
> ${OUTFILE}_${SERVER} #cleanup after ourselves
>
> done
>
>
>
> exit 0
>
>
> -----end script----
>
> ./edy gasparini - Internet Security Manager
>
References:
|
|
