Bash getting a specific section from a file that spans multiple lines

I am using bash on Ubuntu server. My goal is to get the lines from specific section. I can't think of what tool to use to accomplish this (awk, grep -P, sed, ???). It is the specific section that I need help in getting I can grep for the IP:PORT part while looping the section afterwards. Any help would be appreciated.

Lines I need:

backend mc-services [FROM_THIS_SECTION_ONLY]
  server SRV-003 [IP]:[PORT] [ANYTHING TRAILING]

Example file:

backend md-demo
  mode http
  balance roundrobin

  server SRV-001 192.168.2.11:8783 check inter 5s rise 5 fall 5 maxconn 30  

backend mc-services
  mode http
  balance roundrobin

  server SRV-003 192.168.2.13:8784 check inter 5s rise 5 fall 5 maxconn 30  
  server SRV-004 192.168.2.13:8785 check inter 5s rise 5 fall 5 maxconn 30
  server SRV-003 192.168.2.14:8784 check inter 5s rise 5 fall 5 maxconn 30
  server SRV-004 192.168.2.14:8785 check inter 5s rise 5 fall 5 maxconn 30

backend mc-services-int
  mode http
  balance roundrobin

  server SRV-009 192.168.2.11:8789 check inter 5s rise 5 fall 5 maxconn 30

Solution

I almost always recommend sed for layered conditionally applied regexes. To just grab those lines -

$: sed -En '/^backend mc-services *$/,/^backend/ { /^ +server /p }' file
  server SRV-003 192.168.2.13:8784 check inter 5s rise 5 fall 5 maxconn 30
  server SRV-004 192.168.2.13:8785 check inter 5s rise 5 fall 5 maxconn 30
  server SRV-003 192.168.2.14:8784 check inter 5s rise 5 fall 5 maxconn 30
  server SRV-004 192.168.2.14:8785 check inter 5s rise 5 fall 5 maxconn 30

if you just wanted to knock off the first two fields -

$: sed -En '/^backend mc-services *$/,/^backend/ { /^ +server /{ s/^ +server [^ ]+ //; p; } }' file
192.168.2.13:8784 check inter 5s rise 5 fall 5 maxconn 30
192.168.2.13:8785 check inter 5s rise 5 fall 5 maxconn 30
192.168.2.14:8784 check inter 5s rise 5 fall 5 maxconn 30
192.168.2.14:8785 check inter 5s rise 5 fall 5 maxconn 30

Purely as an exercise - don't do this without a reason, it's considerably less effecient, though if your dataset is small enough it won't matter - I include an all-bash script to do the same.

#! /bin/bash
shopt -s extglob
while IFS= read -r line
do case "$line" in
   'backend mc-services')  parsing=1; continue        ;;
   backend*) ((parsing)) && exit ||   continue        ;;
   +( )server\ *) ((parsing))    ||   continue        ;
                  read -r x x addr x <<< "$line"      ;
                  echo "$addr $x"                     ;;
   *)                                 continue        ;;
   esac
done < file

Executed -

$: ./script
192.168.2.13:8784 check inter 5s rise 5 fall 5 maxconn 30
192.168.2.13:8785 check inter 5s rise 5 fall 5 maxconn 30
192.168.2.14:8784 check inter 5s rise 5 fall 5 maxconn 30
192.168.2.14:8785 check inter 5s rise 5 fall 5 maxconn 30