Having an input for a variable, dynamically create values for another variable using a static list

Having an input for a variable, dynamically create values for another variable using a static list

I have a set of hostnames, I would want to assign these hostnames to the variable SERVERS or CLIENTS. When the hostname from the given list is selected for SERVERS it should get removed from the CLIENTS variable.

Note:- variable SERVER might have more than one hostname

> LIST_OF_NODEScn01,cn02,cn03,cn04,cn05,cn06,cn07,cn08,cn09,cn10
INPUT=cn01
>
OUTPUT:-
>
> SERVERS=cn01
> CLIENTS=cn02,cn03,cn04,cn05,cn06,cn07,cn08,cn09,cn10
> (echo "server $SERVERS"; echo "client $CLIENTS") > cn01.cmd
>
INPUT=cn02
>
OUTPUT:-
>
> SERVERS=cn02
> CLIENTS=cn01,cn03,cn04,cn05,cn06,cn07,cn08,cn09,cn10
> (echo "server $SERVERS"; echo "client $CLIENTS") > cn02.cmd
>
> SERVERS=cn01,cn02
> CLIENTS=cn03,cn04,cn05,cn06,cn07,cn08,cn09,cn10
> (echo "server $SERVERS"; echo "client $CLIENTS") > cn01_cn02.cmd

INPUT:- Values for the SERVERS for example

If cn03, cn07 is provided as input then

> SERVERS=cn03,cn07
> CLIENTS=cn01,cn02,cn04,cn05,cn06,cn08,cn09,cn10
> (echo "server $SERVERS"; echo "client $CLIENTS") > cn03_cn07.cmd

I tried the following

for i in cn{01..10}; do 
   echo $i
   sed "s/$i//g" nodes.txt | sed 's/,,/,/g'| sed 's/,*$//g' | sed 's/^,//g' 
done

If I understood your requirement correctly, could you please try following.

cat script.ksh
echo "Enter cn values in form of cn01 or cn02 etc here...."
read value

SERVERS="$value"
LIST_OF_NODES="cn01,cn02,cn03,cn04,cn05,cn06,cn07,cn08,cn09,cn10"   

awk -v servers="$SERVERS" -v clients="$LIST_OF_NODES" '
BEGIN{
  s1=","
  split(servers,array,",")
  for(i in array){
    gsub(s1 array[i] s1,s1,clients)
    gsub("^"array[i] s1,"",clients)
    gsub(s1 array[i]"$","",clients)
  }
  gsub(/,/,"_",servers)
  print clients > servers".cmd"
}'

Created above script. Now following is when we run the script:

./file.ksh
Enter cn values in form of cn01 or cn02 etc here....
cn04,cn07

Output file named cn04_cn07.cmd will be created.

Assumptions:

• nodes.txt contains list of all possible nodes and, initially, all nodes are considered 'client' nodes
• nodes.txt entries may reside on one or more lines, and if multiple nodes are on one line they are delimited by a comma (,)
• input to the script is one or more node names, with multiple nodes delimited by a comma (,)

Sample nodes.txt contents:

$ cat nodes.txt
cn01,cn02,cn03,cn04,cn05,cn06,cn07,cn08,cn09,cn10
cn11,cn12

We'll look at an awk solution to strip 'SERVER' nodes from the 'CLIENT' list of nodes. We'll start by loading all nodes (nodes.txt) into a clients[] array. For the server nodes we'll delete the nodes from the clients[] array. And finally we'll print what's left in the clients[] array.

Example script:

$ cat split_nodes 
SERVERS=                                      # store out input node(s)

outfile=${SERVERS//,/_}.cmd                       # generate output file by replacing ',' with '_'
echo "output file: ${outfile}"

echo "server ${SERVERS}" > "${outfile}"           # dump our 'server' line to ${outfile}

awk '
BEGIN   { FS=RS="[,\n\r]" }                       # input field/record delimiter is "," or "\n" or "\r"

FNR==NR { clients[$1]=1; next}                    # FNR==NR => first file; use node names to populate associative array where index==node name
        { delete clients[$1] }                    # at this point we are processing 2nd file; delete (server) node from client[] array

END     { pfx="client "                           # first part of output is the string "client "
          n=asorti(clients,sorted)                # sort our client nodes by name
          for ( i=1; i<=n; i++ )                  # loop through list of indexes from the sorted array
              { printf "%s%s", pfx, sorted[i]     # print prefix and node name
                pfx=","                           # for array items 2+ change prefix to ","
              }
          printf "\n"                             # terminate output with a line feed
        }
' nodes.txt <(echo "${SERVERS}") >> "${outfile}"  # dump output to our new file

NOTE: <(echo "${SERVERS}") allows us to pass the contents of the variable ${SERVERS} to the awk script in the form of a file.

Running the script with some different input parameters

$ split_nodes cn01
output file: cn01.cmd

$ cat cn01.cmd
server cn01
client cn02,cn03,cn04,cn05,cn06,cn07,cn08,cn09,cn10,cn11,cn12

$ split_nodes cn03,cn07
output file: cn03_cn07.cmd

$ cat cn03_cn07.cmd
server cn03,cn07
client cn01,cn02,cn04,cn05,cn06,cn08,cn09,cn10,cn11,cn12

The script could be modified a good bit to take into consideration different formats for the input (eg, does the input 'string' include any white space, etc), but for now we'll assume the user provides input similar to what's listed in the question.