bash scripting - how to loop through results called within a function - bash

Question

I have a function that gives me a list of IPs and for each IP in my list, I want to run a query. The problem I'm having is its only looping through (1) of the results and not the rest.
getPartition ()
{
_knife=$(which knife);
_grep=$(which grep);
_awk=$(which awk);
cd ~/home/foo/.chef
local result=$(${_knife} search "chef_environment:dev AND role:myapp AND ec2_region:us-east-1" | ${_grep} IP | ${_awk} '{ print $2 }');
read -a servers <<< $result;
echo "Checking ${#servers[#]} servers";
for i in ${servers[#]};
do
local host='10.1.2.123'
local db='mystate'
_mongo=$(which mongo);
echo -n "$i";
local exp="db.foobarcluster_servers.find(
{\"node_host\":\"${i}\",\"node_type\":\"PROCESS\",\"region\":\"us-east-1\",\"status\":\"ACTIVE\"},{\"partition_range_start\":1,\"partition_range_end\":1, _id:0}).pretty();";
${_mongo} ${host}/${db} --eval "$exp" | grep -o -e "{[^}]*}";
done
}
So, I tried using for, but its only running the query for (1) of the (5) hosts listed.
I can see in my output for result that the list of IPs look like this:
+ local 'result=10.8.3.34
10.8.2.161
10.8.3.514
10.8.4.130
10.8.2.173'
So, I'm just returning results for (1) of the IPs it should be (5) of them because I have 5 IPs:
+ read -a servers
+ echo 'Checking 1 servers'
Checking 1 servers
+ for i in ${servers[#]}
+ local host=10.1.2.130
+ local db=mystate
++ which mongo
+ _mongo=/usr/local/bin/mongo
+ echo -n 10.8.3.34
10.8.3.34+ local 'exp=db.foobarcluster_servers.find(
{"node_host":"10.8.3.34","node_type":"PROCESS","region":"us-east-1","status":"ACTIVE"},{"partition_range_start":1,"partition_range_end":1, _id:0}).pretty();'
+ /usr/local/bin/mongo 10.8.3.34/mystate --eval 'db.foobarcluster_servers.find(
{"node_host":"10.8.3.34","node_type":"PROCESS","region":"us-east-1","status":"ACTIVE"},{"partition_range_start":1,"partition_range_end":1, _id:0}).pretty();'
+ grep -o -e '{[^}]*}'
{ "partition_range_start" : 31, "partition_range_end" : 31 }
+ set +x
Results:
{ "partition_range_start" : 31, "partition_range_end" : 31 }
I'm expecting:
{ "partition_range_start" : 31, "partition_range_end" : 31 }
{ "partition_range_start" : 32, "partition_range_end" : 32 }
{ "partition_range_start" : 33, "partition_range_end" : 33 }
{ "partition_range_start" : 34, "partition_range_end" : 34 }
{ "partition_range_start" : 35, "partition_range_end" : 35 }
How do I effectively loop through my IPs? Did I set up result properly as a variable to hold that list of IPs?

Answer 1

Good idea using set -x - another good debugging tactic (that also makes reading set -x easier) would be to comment out parts that aren't relevant to the issue (e.g. make the for loop simply print its iterations, hard-code the value of result, etc.) to try to narrow down the issue.
If I try to replicate what you're doing myself:
demo() {
local result='10.8.3.34
10.8.2.161
10.8.3.514
10.8.4.130
10.8.2.173'
read -a servers <<< $result
echo "Checking ${#servers[#]} servers"
for i in ${servers[#]}; do
echo "$i"
done
}
Which outputs (with set -x):
$ demo
+ demo
+ local 'result=10.8.3.34
10.8.2.161
10.8.3.514
10.8.4.130
10.8.2.173'
+ read -a servers
+ echo 'Checking 5 servers'
Checking 5 servers
+ for i in '${servers[#]}'
+ echo 10.8.3.34
10.8.3.34
+ for i in '${servers[#]}'
+ echo 10.8.2.161
10.8.2.161
+ for i in '${servers[#]}'
+ echo 10.8.3.514
10.8.3.514
+ for i in '${servers[#]}'
+ echo 10.8.4.130
10.8.4.130
+ for i in '${servers[#]}'
+ echo 10.8.2.173
10.8.2.173
In other words, the code you shared appears to be working as expected. Perhaps there's a typo you corrected while transcribing?
A key thing to note (per help read) is that read "Reads a single line from the standard input ... the line is split into fields as with word splitting". In other words, a multi-line input does not all get read by a call to read, only the first line does. We can test this by tweaking the demo function above to use:
read -a servers <<< "$result"
Which causes the output you describe:
$ demo
+ demo
+ local 'result=10.8.3.34
10.8.2.161
10.8.3.514
10.8.4.130
10.8.2.173'
+ read -a servers
+ echo 'Checking 1 servers'
Checking 1 servers
+ for i in '${servers[#]}'
+ echo 10.8.3.34
10.8.3.34
So that's likely the source of your issue - by quoting $result (which generally is a good idea) read respects the newlines separating the elements, and stops reading after it sees the first one.
Instead use the readarray command, which has more sane behavior for tasks like this. It will "read lines from a file into an array variable", rather than stopping after the first line.
You can then skip the indirection of writing to result, as well, and just pipe directly into readarray:
readarray -t servers < <(
${_knife} search "chef_environment:dev AND role:myapp AND ec2_region:us-east-1"
| ${_grep} IP | ${_awk} '{ print $2 }')

Answer 2

it's because read reads only until first input line delimiter "\n"
adding option -d '', reads until end of input
result=serv1$'\n'serv2$'\n'serv3
read -a servers <<< $result
printf "<%s>\n" "${servers[#]}"
read -d '' -a servers <<< $result
printf "<%s>\n" "${servers[#]}"
There's also readarray builtin which can be used to read an array
readarray -t servers <<< $result
printf "<%s>\n" "${servers[#]}"
-t to remove the newlines for each element of the array

Answer 3

The problem is that…
read -a servers <<< $result
is only reading the first line into the array
Change that line to…
servers=( $result )
This converts every whitespace-delimited value in $result into an array element. Effectively servers=( <ip> <ip> <ip> <ip> <ip> )