awk to extract a block of text

Assuming there are only indented lines with 4 whitespace characters after the start, you can make sure that there is at least a single indented line present, or else do not print anything.

awk '
  /^\s{2}Board Info/ {
    start = 1; buffer = $0; indent = 0; next
  }
  start && /^\s{4}\S/ {
    buffer = buffer == "" ? $0 : buffer "\n" $0;
    indent = 1;
  }
  start && /^\s{2}\S/ {
    if (indent) {
      print buffer "\n" $0;
    }
    start = 0;
  }
' dump.txt

Not sure if you want to print the start and the closing line, but you can omit printing them if you want by not adding them to the buffer.

You could change these lines:

start = 1; buffer = ""; indent = 0; next

And:

print buffer;

This awk solution should work with any version of awk:

awk '/^  [^[:blank:]]/ { blk = index($0, "Board Info:"); next } blk' dump.txt
    Manufacturer: "Dell Inc."
    Product: "0X3D66"
    Version: "A02"
    Serial: "..CN7016343F00IE."

Explanation:

• Regular expression /^ [^[:blank:]]/ matches a line with 2 spaces at the start followed by any non-whitespace character.
• blk = index($0, "Board Info:"): sets value of the flag blk to 1 or 0 depending on the condition whether line contains Board Info: or not.
• blk in the end prints a line if blk is 1

Using any awk:

$ awk '/^  [^ ]/{ f=(/Board Info/); next } f' dump.txt
    Manufacturer: "Dell Inc."
    Product: "0X3D66"
    Version: "A02"
    Serial: "..CN7016343F00IE."

The awk script says in pseudo-code:

' = WHILE lines to be read into $0 DO:
    /^  [^ ]/ { = IF the current line starts with 2 blanks then a non-blank THEN
        f=(/Board Info/) = IF the current line contains "Board Info" THEN
                               set f=1
                           ELSE
                               set f=0
                           ENDIF
        next = stop processing the current line
    } = ENDIF
    f = IF f is non-zero THEN
            print the current record
        ENDIF
' = DONE

so if a line starts with only 2 blanks and contains Board Info then f gets set to 1 and awk skips to the next line. The next 4 lines start with more than 2 blanks so the /^ [^ ]/ can't be true so awk just evaluates f which is still 1 and so the lines after Board Info get printed. Next time we sas a line that starts with only 2 blanks it doesn;t contain Board Info, it contains Chassis Info instead, so f is set to 0 and no subsequent lines are printed. Regarding your original code:

awk '/^\s{2}Board Info/,/^\s{2}[^B ]/' dump.txt

1. Using range expressions (/start/,/end/) is usually trickier to get right in general and always harder to enhance than using a flag (f in my code), often leading to duplicate code or other bad software. See Is a /start/,/end/ range expression ever useful in awk? for more info.
2. Only GNU awk would recognize \s as shorthand for [[:space:]] so that makes your code non-portable but both constructs are unnecessary anyway when your spaces are all just blank chars.

A minor tweak of FS is all you need for a printing 1/0 indicator without using pattern ranges :

echo '
  Board Info: #512
    Manufacturer: "Dell Inc."
    Product: "0X3D66"
    Version: "A02"
    Serial: "..CN7016343F00IE."
  Chassis Info: #768' |

awk '^ < NF ?  = !! index($!NF, " Board Info:") : ! < NF' FS='^  [A-Z]'

     1    Board Info: #512
     2      Manufacturer: "Dell Inc."
     3      Product: "0X3D66"
     4      Version: "A02"
     5      Serial: "..CN7016343F00IE."

I know you said awk, but would you consider Perl?

Sometimes Perl is a bit easier when there are newlines in the solution, because Perl can grab the entire file at once. Then sort through the output using a regex. I created a sample output file with two Board Info: sections with some sample text before, after, and in between them.

The regular expression works like this, when you see Board Info: line, backreference everything non-greedy until you see a newline, and two spaces followed by a non space character, then close backreference at the end of the line.

Regex looks like this...

/( {2}Board Info:[\w\W]?\n {2}\S.\n)/g

I'm not sure if you want the last line or not. If you don't want it, change the regex to close the backreference before the closing section, like this...

/( {2}Board Info:[\w\W]?\n) {2}\S.\n/g

Here is the code...

#!/usr/bin/perl -w
my @boardInfo;
undef $/; #grab entire file at once because there are newlines
while(){ #while loop goes exactly once per file
  while(/( {2}Board Info:[\w\W]?\n {2}\S.\n)/g){ #when you see Board Info: line, backreference everything non-greedy
    push(@boardInfo,$1);                            #until you see a newline, and two spaces followed by a non space character,
  }                                                  #then close backreference at the end of the line
}
my $size = @boardInfo;
print "found $size \"Board Info:\" sections\n\n";
for(0 .. $#boardInfo){
  print "match $:\n$boardInfo[$]\n";
}
print "\n";

Output looks like this...

$ perl extract.block.pl extract.block.txt 
found 2 "Board Info:" sections
match 0:
  Board Info: #512
    Manufacturer: "Dell Inc."
    Product: "0X3D66"
    Version: "A02"
    Serial: "..CN7016343F00IE."
  Chassis Info: #768
match 1:
  Board Info: #512
    Manufacturer: "Dell Inc."
    Product: "0X3D66"
    Version: "A02"
    Serial: "..CN7016343F00IE."
  Other Info: #768

Golfed at 61 characters

perl -0777 -ne 'while(/( {2}Board Info:[\w\W]?\n {2}\S.\n)/g){print "$1\n"}' extract.block.txt

between: { ^^ \s\s "Board Info:" }  { ^^ \s\s \S+ }
  regexp: ^^ \s\s\s\s .*
end:
code:  $l {
      say $l;
    }
    say "===";
}
RAKU

  Board Info: #512
      Manufacturer: "Dell Inc."
      Product: "0X3D66"
      Version: "A02"
      Serial: "..CN7016343F00IE."
  Chassis Info: #768
  Board Info: #513
      Manufacturer: "Dell2 Inc."
      Product: "0X3D68"
      Version: "A03"
      Serial: "..CN7016346F00IE."
  bla bla bla: #769

[task check]
stdout match (r)  True
      Manufacturer: "Dell Inc."
      Product: "0X3D66"
      Version: "A02"
      Serial: "..CN7016343F00IE."
===
      Manufacturer: "Dell2 Inc."
      Product: "0X3D68"
      Version: "A03"
      Serial: "..CN7016346F00IE."
===