Use awk in Perl to parse everything between two strings - perl

Question

I have a huge pile of log files constantly being updated on HP-UX server.I have created the Perl code to find out the name of log file in which the string i'm using resides inside.
Perl gets the file name using split and passes it into a variable.Using the userinput i create the start and stop strings as two variables.Such as:
my $ssh = Net::OpenSSH->new($host, user => $user,
master_opts => [ -o => 'NumberOfPasswordPrompts=1',
-o => 'PreferredAuthentications=keyboard-interactive,password'],
login_handler => \&login_handler);
$ssh-> error and die "Unable to connect" . $ssh->error;
my $output=$ssh->capture("grep .$userinput1. /app/bea/user_projects/domains/granite/om_ni.log*");
my $array = (split ":", $output)[0];
print "$array"."\n";
[EDIT]: As you guys requested,above is the beginning of how the $array got filled in.Below is where the awk sequence starts:
my $a= "INFO - $userinput1";print $a;
my $b= "INFO - ProcessNode terminated... [$userinput1]";print $b;
Using the awk as part of ssh capture command,it will search through the whole log file and capture every line between the string $a and string $b,then get everything inside another array.Such as:
my $output2=$ssh->capture("awk -v i=$array '$a,$b' i");
Here $array is where the log file's full path is held and it work completely fine as a passing variable.
I tried using the awk without -v parameter as well,didn't matter at all.
[EDIT 2]:this is the result of print "$array"."\n";
/app/bea/user_projects/domains/granite/om_ni.log.2
When I run the perl script,I get the result:
INFO - 28B26AD1-E959-4F5F-BD89-A7A6E601BE18INFO - ProcessNode terminated... [28B26AD1-E959-4F5F-BD89-A7A6E601BE18] syntax error The source line is 1.
The error context is
INFO - 28B26AD1-E959-4F5F-BD89-A7A6E601BE18,INFO - ProcessNode >>> terminated. <<< .. [28B26AD1-E959-4F5F-BD89-A7A6E601BE18]
awk: Quitting
The source line is 1.
Error pointing at the "terminate" word somehow but even when I use escape characters all over the strings,it just doesn't care and returns the same error.
Any help on this issue is highly appreciated.Thanks a lot in advance.

Answer 1

While I don't really know awk, the way you are invoking it does not appear to be correct. Here is the manual for awk on HP-UX.
The part in single quotes ($a,$b) should be the program. However, you are passing it two text strings, which are not even quoted to separate them. This is not a valid awk program; hence the syntax error.
I think what you want is something like '/$a/, /$b/' for the program (but again, I am not an awk expert).
Also, you are setting the filename to variable i, then using i in the place of the filename when you invoke the command. I don't know why you are doing this, and I don't think it will even work to use a variable in the filename. Just use $array (which you should rename to something like $file for clarity) in the filename position.
So your whole command should look something like:
"awk '/$a/,/$b/' $file"
In this single command, you are dealing with three different tools: Perl, SSH, and awk. This is very hard to debug, because if there is a problem, it can be hard to tell where the problem is. It is essential that you break down the task into smaller parts in order to get something like this working.
In this case, that means that you should manually SSH into the server and play around with awk until you get the command right. Only when you are sure that you have the awk command right should you try incorporating it into Perl. It will be much easier if you break down the task in that way.