solr / tomcat7 doesn't come back up after a crash - unix

Question

I keep getting issues with Solr crashing on my server. Its hardly a busy site, so I'm baffled as to why it keeps doing it.
Anyway, as an intermediary - I'm written a shell script that runs on a cron as root:
#!/bin/bash
declare -a arr=(tomcat7 nginx mysql);
for i in "${arr[#]}"
do
echo "Checking $i"
if (( $(ps -ef | grep -v grep | grep $i | wc -l) > 0 ))
then
echo "$i is running!!!"
else
echo "service $i start\n"
service $i start
fi
done
# re-run, but this time do a restart if its still not going!
for i in "${arr[#]}"
do
echo "Checking $i"
if (( $(ps -ef | grep -v grep | grep $i | wc -l) > 0 ))
then
echo "$i is running!!!"
else
service $i restart
fi
done
..then this cron (as root)
*/5 * * * * bash /root/script-checks.sh
The cron itself seems to run just fine:
Checking tomcat7
service tomcat7 start\n
Checking nginx
nginx is running!!!
Checking mysql
mysql is running!!!
Checking tomcat7
Checking nginx
nginx is running!!!
Checking mysql
mysql is running!!!
...and Tomcats status seems ok:
root#domain:~# service tomcat7 status
â tomcat7.service - LSB: Start Tomcat.
Loaded: loaded (/etc/init.d/tomcat7)
Active: active (exited) since Mon 2016-03-21 06:33:28 GMT; 4 days ago
Process: 2695 ExecStart=/etc/init.d/tomcat7 start (code=exited, status=0/SUCCESS)
Warning: Journal has been rotated since unit was started. Log output is incomplete or unavailable.
...yet my script, can't connect to Solr:
Could not parse JSON response: malformed JSON string, neither tag, array, object, number, string or atom, at character offset 0 (before "Can't connect to loc...") at /srv/www/domain.net/www/cgi-bin/admin/WebService/Solr/Response.pm line 42. Can't connect to localhost:8080 Connection refused at /usr/share/perl5/LWP/Protocol/http.pm line 49.
If I manually run a "restart":
service tomcat7 restart
...it then starts working again. Its almost like the 2nd part in my shell script isn't working.
Any suggestions?
My Solr versions are as follows:
Solr Specification Version: 3.6.2.2014.10.31.18.33.47
Solr Implementation Version: 3.6.2 debian - pbuilder - 2014-10-31 18:33:47
Lucene Specification Version: 3.6.2
UPDATE: I've read that sometimes updating the maxThreads can help with crashes, so I've changed it to 10,000:
<Connector port="8443" protocol="org.apache.coyote.http11.Http11Protocol"
maxThreads="10000" SSLEnabled="true" scheme="https" secure="true"
clientAuth="false" sslProtocol="TLS" />
I guess time will tell, to see if this fixes the issue.

Answer 1

Ok, well I never got to the bottom of why it wouldn't restart... but I have worked out why it was crashing. Before, we had it on 2048mb RAM Linode server, but when we moved over to Apache2, I setup a 1024Mb server, and was going to upgrade it to 2048mb one we had it all working. However, we put it live - but I forgot to update it to the 2048mb server, so Nginx/Apache2/Tomcat/MySQL etc, were all trying to run on a pretty slow server.
We found that Solr was dying with an OOM (out of memory) error, which is what gave us the clue.
Hopefully this helps someone else, who may come across this.