Minimal web server using netcat - bash

I'm trying to set up a minimal web server using netcat (nc). When the browser calls up localhost:1500, for instance, it should show the result of a function (date in the example below, but eventually it'll be a python or c program that yields some data).
My little netcat web server needs to be a while true loop in bash, possibly as simple as this:
while true ; do echo -e "HTTP/1.1 200 OK\n\n $(date)" | nc -l -p 1500 ; done
When I try this the browser shows the currently available data during the moment when nc starts. I want the browser displays the data during the moment the browser requests it, though. How can I achieve this?

Try this:
while true ; do nc -l -p 1500 -c 'echo -e "HTTP/1.1 200 OK\n\n $(date)"'; done
The -cmakes netcat execute the given command in a shell, so you can use echo. If you don't need echo, use -e. For further information on this, try man nc. Note, that when using echo there is no way for your program (the date-replacement) to get the browser request. So you probably finally want to do something like this:
while true ; do nc -l -p 1500 -e /path/to/yourprogram ; done
Where yourprogram must do the protocol stuff like handling GET, sending HTTP 200 etc.

I had the problem where I wanted to return the result of executing a bash command:
$ while true; do { echo -e 'HTTP/1.1 200 OK\r\n'; sh test; } | nc -l 8080; done
NOTE:
This command was taken from: http://www.razvantudorica.com/08/web-server-in-one-line-of-bash
This executes a bash script and returns the result to a browser client connecting to the server running this command on port 8080.
My script does this:
$ nano test
#!/bin/bash
echo "************PRINT SOME TEXT***************\n"
echo "Hello World!!!"
echo "\n"
echo "Resources:"
vmstat -S M
echo "\n"
echo "Addresses:"
echo "$(ifconfig)"
echo "\n"
echo "$(gpio readall)"
and my web browser is showing
************PRINT SOME TEXT***************
Hello World!!!
Resources:
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 0 0 314 18 78 0 0 2 1 306 31 0 0 100 0
Addresses:
eth0 Link encap:Ethernet HWaddr b8:27:eb:86:e8:c5
inet addr:192.168.1.83 Bcast:192.168.1.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:27734 errors:0 dropped:0 overruns:0 frame:0
TX packets:26393 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:1924720 (1.8 MiB) TX bytes:3841998 (3.6 MiB)
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:65536 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
GPIOs:
+----------+-Rev2-+------+--------+------+-------+
| wiringPi | GPIO | Phys | Name | Mode | Value |
+----------+------+------+--------+------+-------+
| 0 | 17 | 11 | GPIO 0 | IN | Low |
| 1 | 18 | 12 | GPIO 1 | IN | Low |
| 2 | 27 | 13 | GPIO 2 | IN | Low |
| 3 | 22 | 15 | GPIO 3 | IN | Low |
| 4 | 23 | 16 | GPIO 4 | IN | Low |
| 5 | 24 | 18 | GPIO 5 | IN | Low |
| 6 | 25 | 22 | GPIO 6 | IN | Low |
| 7 | 4 | 7 | GPIO 7 | IN | Low |
| 8 | 2 | 3 | SDA | IN | High |
| 9 | 3 | 5 | SCL | IN | High |
| 10 | 8 | 24 | CE0 | IN | Low |
| 11 | 7 | 26 | CE1 | IN | Low |
| 12 | 10 | 19 | MOSI | IN | Low |
| 13 | 9 | 21 | MISO | IN | Low |
| 14 | 11 | 23 | SCLK | IN | Low |
| 15 | 14 | 8 | TxD | ALT0 | High |
| 16 | 15 | 10 | RxD | ALT0 | High |
| 17 | 28 | 3 | GPIO 8 | ALT2 | Low |
| 18 | 29 | 4 | GPIO 9 | ALT2 | Low |
| 19 | 30 | 5 | GPIO10 | ALT2 | Low |
| 20 | 31 | 6 | GPIO11 | ALT2 | Low |
+----------+------+------+--------+------+-------+

Add -q 1 to the netcat command line:
while true; do
echo -e "HTTP/1.1 200 OK\n\n $(date)" | nc -l -p 1500 -q 1
done

The problem you are facing is that nc does not know when the web client is done with its request so it can respond to the request.
A web session should go something like this.
TCP session is established.
Browser Request Header: GET / HTTP/1.1
Browser Request Header: Host: www.google.com
Browser Request Header: \n #Note: Browser is telling Webserver that the request header is complete.
Server Response Header: HTTP/1.1 200 OK
Server Response Header: Content-Type: text/html
Server Response Header: Content-Length: 24
Server Response Header: \n #Note: Webserver is telling browser that response header is complete
Server Message Body: <html>sample html</html>
Server Message Body: \n #Note: Webserver is telling the browser that the requested resource is finished.
The server closes the TCP session.
Lines that begin with "\n" are simply empty lines without even a space and contain nothing more than a new line character.
I have my bash httpd launched by xinetd, xinetd tutorial. It also logs date, time, browser IP address, and the entire browser request to a log file, and calculates Content-Length for the Server header response.
user#machine:/usr/local/bin# cat ./bash_httpd
#!/bin/bash
x=0;
Log=$( echo -n "["$(date "+%F %T %Z")"] $REMOTE_HOST ")$(
while read I[$x] && [ ${#I[$x]} -gt 1 ];do
echo -n '"'${I[$x]} | sed -e's,.$,",'; let "x = $x + 1";
done ;
); echo $Log >> /var/log/bash_httpd
Message_Body=$(echo -en '<html>Sample html</html>')
echo -en "HTTP/1.0 200 OK\nContent-Type: text/html\nContent-Length: ${#Message_Body}\n\n$Message_Body"
To add more functionality, you could incorporate.
METHOD=$(echo ${I[0]} |cut -d" " -f1)
REQUEST=$(echo ${I[0]} |cut -d" " -f2)
HTTP_VERSION=$(echo ${I[0]} |cut -d" " -f3)
If METHOD = "GET" ]; then
case "$REQUEST" in
"/") Message_Body="HTML formatted home page stuff"
;;
/who) Message_Body="HTML formatted results of who"
;;
/ps) Message_Body="HTML formatted results of ps"
;;
*) Message_Body= "Error Page not found header and content"
;;
esac
fi
Happy bashing!

Another way to do this
while true; do (echo -e 'HTTP/1.1 200 OK\r\n'; echo -e "\n\tMy website has date function" ; echo -e "\t$(date)\n") | nc -lp 8080; done
Let's test it with 2 HTTP request using curl
In this example, 172.16.2.6 is the server IP Address.
Server Side
admin#server:~$ while true; do (echo -e 'HTTP/1.1 200 OK\r\n'; echo -e "\n\tMy website has date function" ; echo -e "\t$(date)\n") | nc -lp 8080; done
GET / HTTP/1.1 Host: 172.16.2.6:8080 User-Agent: curl/7.48.0 Accept:
*/*
GET / HTTP/1.1 Host: 172.16.2.6:8080 User-Agent: curl/7.48.0 Accept:
*/*
Client Side
user#client:~$ curl 172.16.2.6:8080
My website has date function
Tue Jun 13 18:00:19 UTC 2017
user#client:~$ curl 172.16.2.6:8080
My website has date function
Tue Jun 13 18:00:24 UTC 2017
user#client:~$
If you want to execute another command, feel free to replace $(date).

I had the same need/problem but nothing here worked for me (or I didn't understand everything), so this is my solution.
I post my minimal_http_server.sh (working with my /bin/bash (4.3.11) but not /bin/sh because of the redirection):
rm -f out
mkfifo out
trap "rm -f out" EXIT
while true
do
cat out | nc -l 1500 > >( # parse the netcat output, to build the answer redirected to the pipe "out".
export REQUEST=
while read -r line
do
line=$(echo "$line" | tr -d '\r\n')
if echo "$line" | grep -qE '^GET /' # if line starts with "GET /"
then
REQUEST=$(echo "$line" | cut -d ' ' -f2) # extract the request
elif [ -z "$line" ] # empty line / end of request
then
# call a script here
# Note: REQUEST is exported, so the script can parse it (to answer 200/403/404 status code + content)
./a_script.sh > out
fi
done
)
done
And my a_script.sh (with your need):
#!/bin/bash
echo -e "HTTP/1.1 200 OK\r"
echo "Content-type: text/html"
echo
date

mkfifo pipe;
while true ;
do
#use read line from pipe to make it blocks before request comes in,
#this is the key.
{ read line<pipe;echo -e "HTTP/1.1 200 OK\r\n";echo $(date);
} | nc -l -q 0 -p 8080 > pipe;
done

Here is a beauty of a little bash webserver, I found it online and forked a copy and spruced it up a bit - it uses socat or netcat I have tested it with socat -- it is self-contained in one-script and generates its own configuration file and favicon.
By default it will start up as a web enabled file browser yet is easily configured by the configuration file for any logic. For files it streams images and music (mp3's), video (mp4's, avi, etc) -- I have tested streaming various file types to Linux,Windows and Android devices including a smartwatch!
I think it streams better than VLC actually. I have found it useful for transferring files to remote clients who have no access beyond a web browser e.g. Android smartwatch without needing to worry about physically connecting to a USB port.
If you want to try it out just copy and paste it to a file named bashttpd, then start it up on the host with $> bashttpd -s
Then you can go to any other computer (presuming the firewall is not blocking inbound tcp connections to port 8080 -- the default port, you can change the port to whatever you want using the global variables at the top of the script). http://bashttpd_server_ip:8080
#!/usr/bin/env bash
#############################################################################
###########################################################################
### bashttpd v 1.12
###
### Original author: Avleen Vig, 2012
### Reworked by: Josh Cartwright, 2012
### Modified by: A.M.Danischewski, 2015
### Issues: If you find any issues leave me a comment at
### http://scriptsandoneliners.blogspot.com/2015/04/bashttpd-self-contained-bash-webserver.html
###
### This is a simple Bash based webserver. By default it will browse files and allows for
### retrieving binary files.
###
### It has been tested successfully to view and stream files including images, mp3s,
### mp4s and downloading files of any type including binary and compressed files via
### any web browser.
###
### Successfully tested on various browsers on Windows, Linux and Android devices (including the
### Android Smartwatch ZGPAX S8).
###
### It handles favicon requests by hardcoded favicon image -- by default a marathon
### runner; change it to whatever you want! By base64 encoding your favorit favicon
### and changing the global variable below this header.
###
### Make sure if you have a firewall it allows connections to the port you plan to
### listen on (8080 by default).
###
### By default this program will allow for the browsing of files from the
### computer where it is run.
###
### Make sure you are allowed connections to the port you plan to listen on
### (8080 by default). Then just drop it on a host machine (that has bash)
### and start it up like this:
###
### $192.168.1.101> bashttpd -s
###
### On the remote machine you should be able to browse and download files from the host
### server via any web browser by visiting:
###
### http://192.168.1.101:8080
###
#### This program requires (to work to full capacity) by default:
### socat or netcat (w/ '-e' option - on Ubuntu netcat-traditional)
### tree - useful for pretty directory listings
### If you are using socat, you can type: bashttpd -s
###
### to start listening on the LISTEN_PORT (default is 8080), you can change
### the port below.
### E.g. nc -lp 8080 -e ./bashttpd ## <-- If your nc has the -e option.
### E.g. nc.traditional -lp 8080 -e ./bashttpd
### E.g. bashttpd -s -or- socat TCP4-LISTEN:8080,fork EXEC:bashttpd
###
### Copyright (C) 2012, Avleen Vig <avleen#gmail.com>
###
### Permission is hereby granted, free of charge, to any person obtaining a copy of
### this software and associated documentation files (the "Software"), to deal in
### the Software without restriction, including without limitation the rights to
### use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of
### the Software, and to permit persons to whom the Software is furnished to do so,
### subject to the following conditions:
###
### The above copyright notice and this permission notice shall be included in all
### copies or substantial portions of the Software.
###
### THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
### IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
### FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
### COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
### IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
### CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
###
###########################################################################
#############################################################################
### CHANGE THIS TO WHERE YOU WANT THE CONFIGURATION FILE TO RESIDE
declare -r BASHTTPD_CONF="/tmp/bashttpd.conf"
### CHANGE THIS IF YOU WOULD LIKE TO LISTEN ON A DIFFERENT PORT
declare -i LISTEN_PORT=8080
## If you are on AIX, IRIX, Solaris, or a hardened system redirecting
## to /dev/random will probably break, you can change it to /dev/null.
declare -a DUMP_DEV="/dev/random"
## Just base64 encode your favorite favicon and change this to whatever you want.
declare -r FAVICON="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"
declare -i DEBUG=1
declare -i VERBOSE=0
declare -a REQUEST_HEADERS
declare REQUEST_URI=""
declare -a HTTP_RESPONSE=(
[200]="OK"
[400]="Bad Request"
[403]="Forbidden"
[404]="Not Found"
[405]="Method Not Allowed"
[500]="Internal Server Error")
declare DATE=$(date +"%a, %d %b %Y %H:%M:%S %Z")
declare -a RESPONSE_HEADERS=(
"Date: $DATE"
"Expires: $DATE"
"Server: Slash Bin Slash Bash"
)
function warn() { ((${VERBOSE})) && echo "WARNING: $#" >&2; }
function chk_conf_file() {
[ -r "${BASHTTPD_CONF}" ] || {
cat >"${BASHTTPD_CONF}" <<'EOF'
#
# bashttpd.conf - configuration for bashttpd
#
# The behavior of bashttpd is dictated by the evaluation
# of rules specified in this configuration file. Each rule
# is evaluated until one is matched. If no rule is matched,
# bashttpd will serve a 500 Internal Server Error.
#
# The format of the rules are:
# on_uri_match REGEX command [args]
# unconditionally command [args]
#
# on_uri_match:
# On an incoming request, the URI is checked against the specified
# (bash-supported extended) regular expression, and if encounters a match the
# specified command is executed with the specified arguments.
#
# For additional flexibility, on_uri_match will also pass the results of the
# regular expression match, ${BASH_REMATCH[#]} as additional arguments to the
# command.
#
# unconditionally:
# Always serve via the specified command. Useful for catchall rules.
#
# The following commands are available for use:
#
# serve_file FILE
# Statically serves a single file.
#
# serve_dir_with_tree DIRECTORY
# Statically serves the specified directory using 'tree'. It must be
# installed and in the PATH.
#
# serve_dir_with_ls DIRECTORY
# Statically serves the specified directory using 'ls -al'.
#
# serve_dir DIRECTORY
# Statically serves a single directory listing. Will use 'tree' if it is
# installed and in the PATH, otherwise, 'ls -al'
#
# serve_dir_or_file_from DIRECTORY
# Serves either a directory listing (using serve_dir) or a file (using
# serve_file). Constructs local path by appending the specified root
# directory, and the URI portion of the client request.
#
# serve_static_string STRING
# Serves the specified static string with Content-Type text/plain.
#
# Examples of rules:
#
# on_uri_match '^/issue$' serve_file "/etc/issue"
#
# When a client's requested URI matches the string '/issue', serve them the
# contents of /etc/issue
#
# on_uri_match 'root' serve_dir /
#
# When a client's requested URI has the word 'root' in it, serve up
# a directory listing of /
#
# DOCROOT=/var/www/html
# on_uri_match '/(.*)' serve_dir_or_file_from "$DOCROOT"
# When any URI request is made, attempt to serve a directory listing
# or file content based on the request URI, by mapping URI's to local
# paths relative to the specified "$DOCROOT"
#
#unconditionally serve_static_string 'Hello, world! You can configure bashttpd by modifying bashttpd.conf.'
DOCROOT=/
on_uri_match '/(.*)' serve_dir_or_file_from
# More about commands:
#
# It is possible to somewhat easily write your own commands. An example
# may help. The following example will serve "Hello, $x!" whenever
# a client sends a request with the URI /say_hello_to/$x:
#
# serve_hello() {
# add_response_header "Content-Type" "text/plain"
# send_response_ok_exit <<< "Hello, $2!"
# }
# on_uri_match '^/say_hello_to/(.*)$' serve_hello
#
# Like mentioned before, the contents of ${BASH_REMATCH[#]} are passed
# to your command, so its possible to use regular expression groups
# to pull out info.
#
# With this example, when the requested URI is /say_hello_to/Josh, serve_hello
# is invoked with the arguments '/say_hello_to/Josh' 'Josh',
# (${BASH_REMATCH[0]} is always the full match)
EOF
warn "Created bashttpd.conf using defaults. Please review and configure bashttpd.conf before running bashttpd again."
# exit 1
}
}
function recv() { ((${VERBOSE})) && echo "< $#" >&2; }
function send() { ((${VERBOSE})) && echo "> $#" >&2; echo "$*"; }
function add_response_header() { RESPONSE_HEADERS+=("$1: $2"); }
function send_response_binary() {
local code="$1"
local file="${2}"
local transfer_stats=""
local tmp_stat_file="/tmp/_send_response_$$_"
send "HTTP/1.0 $1 ${HTTP_RESPONSE[$1]}"
for i in "${RESPONSE_HEADERS[#]}"; do
send "$i"
done
send
if ((${VERBOSE})); then
## Use dd since it handles null bytes
dd 2>"${tmp_stat_file}" < "${file}"
transfer_stats=$(<"${tmp_stat_file}")
echo -en ">> Transferred: ${file}\n>> $(awk '/copied/{print}' <<< "${transfer_stats}")\n" >&2
rm "${tmp_stat_file}"
else
## Use dd since it handles null bytes
dd 2>"${DUMP_DEV}" < "${file}"
fi
}
function send_response() {
local code="$1"
send "HTTP/1.0 $1 ${HTTP_RESPONSE[$1]}"
for i in "${RESPONSE_HEADERS[#]}"; do
send "$i"
done
send
while IFS= read -r line; do
send "${line}"
done
}
function send_response_ok_exit() { send_response 200; exit 0; }
function send_response_ok_exit_binary() { send_response_binary 200 "${1}"; exit 0; }
function fail_with() { send_response "$1" <<< "$1 ${HTTP_RESPONSE[$1]}"; exit 1; }
function serve_file() {
local file="$1"
local CONTENT_TYPE=""
case "${file}" in
*\.css)
CONTENT_TYPE="text/css"
;;
*\.js)
CONTENT_TYPE="text/javascript"
;;
*)
CONTENT_TYPE=$(file -b --mime-type "${file}")
;;
esac
add_response_header "Content-Type" "${CONTENT_TYPE}"
CONTENT_LENGTH=$(stat -c'%s' "${file}")
add_response_header "Content-Length" "${CONTENT_LENGTH}"
## Use binary safe transfer method since text doesn't break.
send_response_ok_exit_binary "${file}"
}
function serve_dir_with_tree() {
local dir="$1" tree_vers tree_opts basehref x
## HTML 5 compatible way to avoid tree html from generating favicon
## requests in certain browsers, such as browsers in android smartwatches. =)
local no_favicon=" <link href=\"data:image/x-icon;base64,${FAVICON}\" rel=\"icon\" type=\"image/x-icon\" />"
local tree_page=""
local base_server_path="/${2%/}"
[ "$base_server_path" = "/" ] && base_server_path=".."
local tree_opts="--du -h -a --dirsfirst"
add_response_header "Content-Type" "text/html"
# The --du option was added in 1.6.0. "/${2%/*}"
read _ tree_vers x < <(tree --version)
tree_page=$(tree -H "$base_server_path" -L 1 "${tree_opts}" -D "${dir}")
tree_page=$(sed "5 i ${no_favicon}" <<< "${tree_page}")
[[ "${tree_vers}" == v1.6* ]]
send_response_ok_exit <<< "${tree_page}"
}
function serve_dir_with_ls() {
local dir="$1"
add_response_header "Content-Type" "text/plain"
send_response_ok_exit < \
<(ls -la "${dir}")
}
function serve_dir() {
local dir="$1"
# If `tree` is installed, use that for pretty output.
which tree &>"${DUMP_DEV}" && \
serve_dir_with_tree "$#"
serve_dir_with_ls "$#"
fail_with 500
}
function urldecode() { [ "${1%/}" = "" ] && echo "/" || echo -e "$(sed 's/%\([[:xdigit:]]\{2\}\)/\\\x\1/g' <<< "${1%/}")"; }
function serve_dir_or_file_from() {
local URL_PATH="${1}/${3}"
shift
URL_PATH=$(urldecode "${URL_PATH}")
[[ $URL_PATH == *..* ]] && fail_with 400
# Serve index file if exists in requested directory
[[ -d "${URL_PATH}" && -f "${URL_PATH}/index.html" && -r "${URL_PATH}/index.html" ]] && \
URL_PATH="${URL_PATH}/index.html"
if [[ -f "${URL_PATH}" ]]; then
[[ -r "${URL_PATH}" ]] && \
serve_file "${URL_PATH}" "$#" || fail_with 403
elif [[ -d "${URL_PATH}" ]]; then
[[ -x "${URL_PATH}" ]] && \
serve_dir "${URL_PATH}" "$#" || fail_with 403
fi
fail_with 404
}
function serve_static_string() {
add_response_header "Content-Type" "text/plain"
send_response_ok_exit <<< "$1"
}
function on_uri_match() {
local regex="$1"
shift
[[ "${REQUEST_URI}" =~ $regex ]] && \
"$#" "${BASH_REMATCH[#]}"
}
function unconditionally() { "$#" "$REQUEST_URI"; }
function main() {
local recv=""
local line=""
local REQUEST_METHOD=""
local REQUEST_HTTP_VERSION=""
chk_conf_file
[[ ${UID} = 0 ]] && warn "It is not recommended to run bashttpd as root."
# Request-Line HTTP RFC 2616 $5.1
read -r line || fail_with 400
line=${line%%$'\r'}
recv "${line}"
read -r REQUEST_METHOD REQUEST_URI REQUEST_HTTP_VERSION <<< "${line}"
[ -n "${REQUEST_METHOD}" ] && [ -n "${REQUEST_URI}" ] && \
[ -n "${REQUEST_HTTP_VERSION}" ] || fail_with 400
# Only GET is supported at this time
[ "${REQUEST_METHOD}" = "GET" ] || fail_with 405
while IFS= read -r line; do
line=${line%%$'\r'}
recv "${line}"
# If we've reached the end of the headers, break.
[ -z "${line}" ] && break
REQUEST_HEADERS+=("${line}")
done
}
if [[ ! -z "{$1}" ]] && [ "${1}" = "-s" ]; then
socat TCP4-LISTEN:${LISTEN_PORT},fork EXEC:"${0}"
else
main
source "${BASHTTPD_CONF}"
fail_with 500
fi

LOL, a super lame hack, but at least curl and firefox accepts it:
while true ; do (dd if=/dev/zero count=10000;echo -e "HTTP/1.1\n\n $(date)") | nc -l 1500 ; done
You better replace it soon with something proper!
Ah yes, my nc were not exactly the same as yours, it did not like the -p option.

If you're using Apline Linux, the BusyBox netcat is slightly different:
while true; do nc -l -p 8080 -e sh -c 'echo -e "HTTP/1.1 200 OK\n\n$(date)"'; done
And another way using printf:
while true; do nc -l -p 8080 -e sh -c "printf 'HTTP/1.1 200 OK\n\n%s' \"$(date)\""; done

while true; do (echo -e 'HTTP/1.1 200 OK\r\nConnection: close\r\n';) | timeout 1 nc -lp 8080 ; done
Closes connection after 1 sec, so curl doesn't hang on it.

Type in nc -h and see if You have -e option available. If yes, You can create a script, for example:
script.sh
echo -e "HTTP/1.1 200 OK\n\n $(date)"
and run it like this:
while true ; do nc -l -p 1500 -e script.sh; done
Note that -e option needs to be enabled at compilation to be available.

I think the problem that all the solution listed doesn't work, is intrinsic in the nature of http service, the every request established is with a different client and the response need to be processed in a different context, every request must fork a new instance of response...
The current solution I think is the -e of netcat but I don't know why doesn't work... maybe is my nc version that I test on openwrt...
with socat it works....
I try this https://github.com/avleen/bashttpd
and it works, but I must run the shell script with this command.
socat tcp-l:80,reuseaddr,fork EXEC:bashttpd &
The socat and netcat samples on github doesn't works for me, but the socat that I used works.

Actually, the best way to close gracefully the connection is to send the Content-Length header like following. Client (like curl will close the connection after receiving the data.
DATA="Date: $(date)";
LENGTH=$(echo $DATA | wc -c);
echo -e "HTTP/1.1 200 OK\nContent-Length: ${LENGTH}\n\n${DATA}" | nc -l -p 8000;

On OSX you can use :
while true; do echo -e "HTTP/1.1 200 OK\n\n $(date)" | nc -l localhost 1500 ; done

Related

pipe and grep in Docker container seems to get blocked

I have a simple bash script as follows that is part of a docker image.
test.sh,
#!/bin/bash
set -e
logit() {
log_date=`date +"%F %T"`
echo "[$log_date][INFO] $1"
}
waitForServerToStart() {
while true; do
logit "Testing .... 1"
netstat -anpt
logit "Testing .... 2"
netstat -anpt | grep tcp
logit "Testing .... 3"
sleep 5
logit "Testing .... 4"
done
}
waitForServerToStart
run.sh,
#!/bin/sh
/test.sh &
# Run forever
while true; do sleep 5; done
Dockerfile,
FROM openjdk:8u191-jre-alpine3.9
COPY files/run.sh /
COPY files/test.sh /
CMD ["/run.sh"]
If I run this container I only get the following output which leads me to believe somehow grep and "pipe" seem to get blocked.
[2019-03-06 11:10:45][INFO] Testing .... 1
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name
tcp 0 0 172.17.0.2:58278 xxx.xxx.xx.xx:443 FIN_WAIT2 -
[2019-03-06 11:10:45][INFO] Testing .... 2
Can someone please shed some light around this ?
It works fine If I comment out netstat -anpt | grep tcp. I would then see the subsequent log lines and it would also continue in the loop.
[2019-03-06 11:25:36][INFO] Testing .... 3
[2019-03-06 11:25:41][INFO] Testing .... 4
This one has me puzzled! But I have a solution for you:
Use awk instead of grep
In test.sh use this instead:
netstat -anpt | awk /tcp/
So that the file looks like this:
#!/bin/bash
set -e
logit() {
log_date=`date +"%F %T"`
echo "[$log_date][INFO] $1"
}
waitForServerToStart() {
while true; do
logit "Testing .... 1"
netstat -anpt
logit "Testing .... 2"
netstat -anpt | awk /tcp/
logit "Testing .... 3"
sleep 5
logit "Testing .... 4"
done
}
waitForServerToStart
For a reason that I cannot explain - grep will not return when reading from the pipe when invoked from the script. I created your container locally, ran it and entered it - and the command netstat -anpt | grep tcp runs just fine and exits. If you replace it with netstat -anpt | cat in your test.sh script, then it will also pass just fine.
I looked all over the place for the someone with an identical issue with grep in a container from the distro you are using, the version etc. - but came up empty handed.
I believe that it may have to do with grep waiting for a EOF character that never lands - but I am not sure.

Jenkins Reload Configuration Removes Changes

Currently using Jenkins at my company, I set up a server for all of our engineers to plug in to, in doing so I made some server management jobs to make my life a little easier. One of them was a config editor to edit the $JENKINS_HOME/config.xml file and trigger a configuration reload to reflect the new changes.
However today when I went to go use that job, the changes we no longer taking effect, nor were they shown when ssh'd into the server and cat-ing the config.xml file.
Did some debugging, made sure that the file contents were being replaced correctly, even threw the checks into the build executor to make sure I knew that everything was correct prior to running the reload-configuration command by double checking md5 sums as the entire content is replaced in my script. I even sleep 15-d before the reload so I could cat the config.xml file and ensure my changes are there, and they always are.
However, as soon as the reload command is run, all of my changes are replaced with what the config contents were just before I made my changes (I also confirmed this from md5 sums of the file in my debugging)
Here's the executor of my job if that helps at all:
$CONFIG_FILE is always $JENKINS_HOME/config.xml
#!/bin/bash
set -o pipefail -e -u -x
cp "$CONFIG_FILE" "$WORKSPACE/config_backup.xml"
printf "Creating an AMI profile with these parameters: \n\n\
Config File: | $CONFIG_FILE \n\
AMI ID: | $AMI_ID \n\
Description: | $DESCRIPTION \n\
Instance Type: | $INSTANCE_TYPE \n\
Security Groups: | $SECURITY_GROUPS \n\
Remote Workspace: | $REMOTE_WORKSPACE \n\
Label(s): | $LABELS \n\
Subnet ID: | $SUBNET_ID \n\
IAM Profile: | $IAM_INSTANCE_PROFILE \n\
Instance Tags: | $TAGS \n\
Executors: | $EXECUTORS \
\n\n\
"
new_xml="$(python "$WORKSPACE/<scriptname removed for security reasons>" \
--file $CONFIG_FILE \
--ami $AMI_ID \
--description $DESCRIPTION \
--type $INSTANCE_TYPE \
--security-groups $SECURITY_GROUPS \
--remote-workspace $REMOTE_WORKSPACE \
--labels $LABELS \
--iam-instance-profile $IAM_INSTANCE_PROFILE \
--subnet-id $SUBNET_ID \
--tags $TAGS \
--executors $EXECUTORS)" || true
if [ -z "$new_xml" ]; then
echo "Ran into an error..."
cat "xml_ami_profile_parser.log"
exit 1
fi
echo "setting new config file content..."
echo "$new_xml" > "$CONFIG_FILE"
echo "config file set!"
CONFIG_MD5="$(md5sum "$CONFIG_FILE" | awk '{print $1}')"
NEW_MD5="$(echo "$new_xml" | md5sum | awk '{print $1}')"
printf "comparing MD5 Sums: \n\
[ $CONFIG_MD5 ] \n\
[ $NEW_MD5 ]\n\n"
if [[ "$CONFIG_MD5" != "$NEW_MD5" ]]; then
echo "Config File ($CONFIG_FILE) was not overwritten successfully. Restoring backup..."
cp "$WORKSPACE/config_backup.xml" "$CONFIG_FILE"
exit 1
fi
# use jenkins api user info
USERNAME="$(cat <scriptname removed for security reasons> | awk '{print $8}')"
PASSWORD="$(cat <scriptname removed for security reasons> | awk '{print $9}')"
curl -X POST -u "$USERNAME:$PASSWORD" "<url removed for security reasons>"
sleep 10
NEW_MD5="$(md5sum "$CONFIG_FILE" | awk '{print $1}')"
printf "comparing MD5 Sums: \n\
[ $CONFIG_MD5 ] \n\
[ $NEW_MD5 ]\n\n"
if [[ "$CONFIG_MD5" != "$NEW_MD5" ]]; then
echo "Config file reverted after reload, marking build as error."
exit 1
fi
Any help at all is greatly appreciated!
EDIT:
Here's the common output now and can't get past it:
setting new config file content...
config file set!
comparing MD5 Sums:
[ 58473de6acbb48b2e273e3395e64ed0f ]
[ 58473de6acbb48b2e273e3395e64ed0f ]
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
comparing MD5 Sums:
[ 58473de6acbb48b2e273e3395e64ed0f ]
[ f521cec2a2e376921995f773522f78e1 ]
Config file reverted after reload, marking build as error.
Build step 'Execute shell' marked build as failure
Finished: FAILURE
For everyone coming to this later, I solved my own problem. Jenkins has it's own failsafe to keep uptime but doesn't give you any notice of it doing so. If you replace a config.xml with something that a plugin can't parse correctly (in my case the Amazon EC2 Plugin) then the plugin tells Jenkins that the config file is bad, and Jenkins will revert to the last correct XML file it was using (usually the one it has in memory).
If this happens to you double check that you aren't using special chars.
the offending code in mine was an output of the tags section including html char converted quotations " -> " and the plugin couldn't parse this. It was solely a difference in:
<tags>
<hudson.plugins.ec2.EC2Tag>
<name>"Email</name>
<value><removed for security reasons>"</value>
</hudson.plugins.ec2.EC2Tag>
<hudson.plugins.ec2.EC2Tag>
<name>"Name</name>
<value><removed for security reasons>"</value>
</hudson.plugins.ec2.EC2Tag>
</tags>
and
<tags>
<hudson.plugins.ec2.EC2Tag>
<name>Email</name>
<value><removed for security reasons></value>
</hudson.plugins.ec2.EC2Tag>
<hudson.plugins.ec2.EC2Tag>
<name>Name</name>
<value><removed for security reasons></value>
</hudson.plugins.ec2.EC2Tag>
</tags>

Reply to IRC PING with PONG in bash script

So far I have a basic IRC Bot which I hope to be able to successfully PRIVMSG myself with but on the server I am testing on it requires a PONG response to PING.
I have absolutely no idea how to get around this. How do I reply to the PING?
My current code:
#!/bin/bash
function ircpreamble {
echo "NICK ${1}"
}
function privmsg {
TARGET=$1
sed -re "s/^(.*)\$/PRIVMSG ${TARGET} :\1/"
}
function delay {
while read LINE; do
sleep 2
echo $LINE
done
}
function messages {
msg=`cat pmmsg.txt`
echo $msg
}
function disconnect {
echo "QUIT goodbye :)"
}
(
nick=`cat randnick.txt`
pms=`cat pmnickname.txt`
ircpreamble "$nick";
messages | privmsg "$pms";
disconnect;
) | delay | nc irc.seersirc.net 6667
You first need to "catch" the server responses. This can be done like:
$ nc irc.seersirc.net 6667 |while read res;do echo "==>$res";done
==>:irc.seersirc.net NOTICE AUTH :*** Looking up your hostname...
==>:irc.seersirc.net NOTICE AUTH :*** Couldn't resolve your hostname; using your IP address instead
Since now you catch the response in a variable, you can filter out these responses and send the appropriate commands back to the server.
The problem is that when running netcat/telnet from a script a simple echo "PONG" is not sending the messages back to server but echo prints the messages locally in your terminal.
To send messages to the server you need either to pipe those messages to netcat in the beginning (the technique you already use) or to use some kind of expect script or to use a file to feed the netcat.
Regarding the PONG response, note that in order pong to be accepted by the server it must include the message sent by the server along with it's ping request.
Your server sends something like PING :D7AA1D1D (different every time) and thus the correct pong response is PONG :D7AA1D1D
As a result you can not just include a pong response in the first messages send in the beginning to netcat, since you don't know what is the ID that your server will sent to you along with the ping request.
This is a working draft script using a file to continuously feed the netcat:
rm .ircbot
touch .ircbot
prmnick="gv"
tail -f .ircbot |nc irc.seersirc.net 6667 |while read res
do
echo "==>$res"
if [[ "$res" == *"Couldn't resolve your hostname; using your IP address instead"* ]];then
sleep 2
echo "NICK gvgv" >>.ircbot
tail -n1 .ircbot #used just to display the last line of the helper file in my screen.
elif [[ "$res" == *"PING"* ]]; then
sleep 2
echo "$res" |sed 's/PING/PONG/' >>.ircbot
tail -n1 .ircbot
sleep 2
echo "USER gvgv 8 * :gvgv " >>.ircbot
tail -n1 .ircbot
sleep 2
echo "PRIVMSG $prmnick : hello from bot" >>.ircbot
tail -n1 .ircbot
fi
done
Tip: By opening a second terminal , you can manually "control" above bot by sending more commands to the .ircbot file (i.e $ echo "JOIN #channel" >>.ircbot) which will be also fed to netcat.
By the way, some web search about bash irc bots will return some useful results.
This is one easy bash script to use as an irc bot: https://github.com/Newbrict/bash-irc-bot/blob/master/bot.sh
Also , i found this useful IRC Over Telnet guide: http://archive.oreilly.com/pub/h/1963
Finally , this is an alternative using the /dev/tcp directly : https://gist.github.com/Wollw/3330337

I want to use parallel-ssh to run a bash script on multiple servers, but it simple prints the echo statements

I have a bash script called sr_run_batch.sh which does super resolution of images. Now I want to do testing on different servers in parallel at the same time. ie. 1 Virtual machine at one given point of time. then 2 virtual machines at one point of time , 3 and then 4.
I tried writing into it the commands
for host in $(cat hosts.txt); do ssh "$host" "$command" >"output.$host"; done
ssh-keygen && for host in $(cat hosts.txt); do ssh-copy-id $host; done
where the file hosts.txt contains the list of servers: username#ip(format) but when I run this, it gives me substitution error
Hence, I tried pssh (parallel-ssh)
pssh -h hosts-file -l username -P $command
command being ./sr_run_batch.sh
but it didn't run, so I modified this to
pssh -h hosts-file -l ben -P -I<./sr_run_batch.sh
But, for some unknown reason, it just prints the echo statements in the code.
here is the code :
NList=(5)
VList=(1)
FList=("input/flower1.jpg" "input/flower2.jpg" "input/flower3.jpg" "input/flower4.jpg")
IList=("320X240" "640X480" "1280X960" "1920X1200")
SList=(2 3)
for VM in ${VList[#]}; do
for ((index=0; index < ${#FList};)) do
file=$FList[$index]
image_size=$IList[$index]
width=`echo $image_size|cut -d "X" -f1`
height=`echo $image_size|cut -d "X" -f2`
for scale_factor in ${SList[#]}; do
for users in ${NList[#]}; do
echo "V: $VM, " "F: $file, " "S: $scale_factor, " "I: $width $height , " "N: $users"
for i in `seq 1 $users` ; do
./sr_run_once.sh $file $width $height $scale_factor &
done
wait
done # for users
done # for scale_factor
done # for index
done # for VM
exit 0
Have you also tried to use pssh with a simple bash-script so see if the communication is set up ok?
$ pssh -h hosts.txt -A -l ben -P -I<./uptime.sh
Warning: do not enter your password if anyone else has superuser
privileges or access to your account.
Password:
10.0.0.67: 11:06:50 up 28 min, 2 users, load average: 0.00, 0.00, 0.00
[1] 11:06:50 [SUCCESS] 10.0.0.67
10.0.0.218: 11:06:50 up 24 min, 2 users, load average: 0.00, 0.05, 0.20
[2] 11:06:50 [SUCCESS] 10.0.0.218

SQUID3 - using multiple auth_param like basic_ncsa_auth & basic_ldap_auth

i tried to setup squid3 with multiple auth_param. Basically, the first choice should be basic_ldap_auth and if this doesnt return OK it should try basic_ncsa_auth with the same values. As far as i know squid doesnt support it however there is the possibility to use "external" ACL
auth_param basic program /usr/lib/squid3/basic_fake_auth
external_acl_type MultAuth %SRC %LOGIN %{Proxy-Authorization} /etc/squid3/multAuth.pl
acl extAuth external MultAuth
my "multAuth.pl"
use URI::Escape;
use MIME::Base64;
$|=1;
while (<>) {
($ip,$user,$auth) = split();
# Retrieve the password from the authentication header
$auth = uri_unescape($auth);
($type,$authData) = split(/ /, $auth);
$authString = decode_base64($authData);
($username,$password) = split(/:/, $authString);
# do the authentication and pass results back to Squid.
$ldap = `/bin/bash auth/ldap.sh`;
if ($ldap == "OK") {
print "OK";
}
$ncsa = `/bin/bash auth/ncsa.sh`;
if ($ncsa == "OK") {
print "OK";
} else {
print "ERR";
}
}
now i am trying to run with ncsa.sh and ldap.sh the "normal" shell command for these auth methods.
./basic_ldap_auth -R -b "dc=domain,dc=de" -D "CN=Administrator,CN=Users,DC=domain,DC=de" -w "password" -f sAMAccountName=%s -h domain.de
user password
and
./basic_ncsa_auth /etc/squid3/users
user password
Therefor i ran:
auth/ncsa.sh
#!/usr/bin/expect
eval spawn [lrange $argv 0 end]
expect ""
send [lindex $argv 1]
send '\r'
expect {
"OK" {
echo OK
exp_continue
}
"ERR" {
echo ERR
exp_continue
}
interact
with
./ncsa.sh "/usr/lib/squid3/basic_ncsa_auth /etc/squid3/users" "user password"
and i generate the following error:
couldn't execute "/usr/lib/squid3/basic_ncsa_auth /etc/squid3/users": no such file or directory
while executing
"spawn {/usr/lib/squid3/basic_ncsa_auth /etc/squid3/users} {user password}"
("eval" body line 1)
invoked from within
"eval spawn [lrange $argv 0 end]"
(file "./ncsa.sh" line 2)
Besides this error, i am not sure how to pass the variables (username & password) forward and i am also not sure how to answer the shell questions like for example the user & pw input for basic_ldap_auth .
Is there a nice way how to solve that? or any other good plan ?
thanks!
FWIW, the following script helped me transition from passwd based to LDAP based authentication.
Contrary to your requirements, my script acts the other way around: It first checks passwd, then LDAP.
#!/usr/bin/env bash
# multiple Squid basic auth checks
# originally posted here: https://github.com/HackerHarry/mSquidAuth
#
# credits
# https://stackoverflow.com/questions/24147067/verify-user-and-password-against-a-file-created-by-htpasswd/40131483
# https://stackoverflow.com/questions/38710483/how-to-stop-ldapsearch1-from-base64-encoding-userpassword-and-other-attributes
#
# requires ldap-utils, openssl and perl
# tested with Squid 4 using a "auth_param basic program /usr/lib/squid/mSquidAuth.sh" line
# authenticate first against squid password file
# if this fails, try LDAP (Active Directory) and also check group membership
# variables
# sLOGFILE=/var/log/squid/mSquidAuth.log
sPWDFILE="/etc/squid/passwd"
sLDAPHOST="ldaps://dc.domain.local:636"
sBASE="DC=domain,DC=local"
sLDS_OPTIONS="-o ldif-wrap=no -o nettimeout=7 -LLL -P3 -x "
sBINDDN="CN=LDAP-read-user,OU=Users,DC=domain,DC=local"
sBINDPW="read-user-password"
sGROUP="Proxy-Users"
# functions
function _grantAccess {
# echo "access granted - $sUSER" >>$sLOGFILE
echo "OK"
}
function _denyAccess {
# echo "access denied - $sUSER" >>$sLOGFILE
echo "ERR"
}
function _setUserAndPass {
local sAuth="$1"
local sOldIFS=$IFS
IFS=' '
set -- $sAuth
IFS=$sOldIFS
# set it globally
sUSER="$1"
sPASS="$2"
}
# loop
while (true); do
read -r sAUTH
sUSER=""
sPASS=""
sSALT=""
sUSERENTRY=""
sHASHEDPW=""
sUSERDN=""
iDNCOUNT=0
if [ -z "$sAUTH" ]; then
# echo "exiting" >>$sLOGFILE
exit 0
fi
_setUserAndPass "$sAUTH"
sUSERENTRY=$(grep -E "^${sUSER}:" "$sPWDFILE")
if [ -n "$sUSERENTRY" ]; then
sSALT=$(echo "$sUSERENTRY" | cut -d$ -f3)
if [ -n "$sSALT" ]; then
sHASHEDPW=$(openssl passwd -apr1 -salt "$sSALT" "$sPASS")
if [ "$sUSERENTRY" = "${sUSER}:${sHASHEDPW}" ]; then
_grantAccess
continue
fi
fi
fi
# LDAP is next
iDNCOUNT=$(ldapsearch $sLDS_OPTIONS -H "$sLDAPHOST" -D "$sBINDDN" -w "$sBINDPW" -b "$sBASE" "(|(sAMAccountName=${sUSER})(userPrincipalName=${sUSER}))" dn 2>/dev/null | grep -cE 'dn::? ')
if [ $iDNCOUNT != 1 ]; then
# user needs a unique account
_denyAccess
continue
fi
# get user's DN
# we need the extra grep in case we get lines back starting with "# refldap" :/
sUSERDN=$(ldapsearch $sLDS_OPTIONS -H "$sLDAPHOST" -D "$sBINDDN" -w "$sBINDPW" -b "$sBASE" "(|(sAMAccountName=${sUSER})(userPrincipalName=${sUSER}))" dn 2>/dev/null | perl -MMIME::Base64 -n -00 -e 's/\n +//g;s/(?<=:: )(\S+)/decode_base64($1)/eg;print' | grep -E 'dn::? ' | sed -r 's/dn::? //')
# try and bind using that DN to check password validity
# also test if that user is member of a particular group
# backslash in DN needs special treatment
if ldapsearch $sLDS_OPTIONS -H "$sLDAPHOST" -D "$sUSERDN" -w "$sPASS" -b "$sBASE" "name=${sGROUP}" member 2>/dev/null | perl -MMIME::Base64 -n -00 -e 's/\n +//g;s/(?<=:: )(\S+)/decode_base64($1)/eg;print' | grep -q "${sUSERDN/\\/\\\\}"; then
_grantAccess
continue
fi
_denyAccess
done

Resources