TR logging: Archive logs using an integer counter

[zrhoffman]

#6450 fixes #6446 by revert the Traffic Router rollover file patterns to the one previously used in log4j.properties with Log4j 1.2, as well as ensuring that traffic_router.log only keeps 1 archive.

---

Which Traffic Control components are affected by this PR?

• Traffic Router
• Automation - Ansible Roles

What is the best way to verify this PR?

1. For convenience, change the value of SizeBasedTriggeringPolicy in log4j2.xml from 200MB and 100MB to 1MB so log file rollovers happen more frequently
2. Fill access.log, access.log.1, traffic_router.log, and traffic_router.log.1 with content so they are over 1MB in size:

   cd /opt/traffic_router/var/log;
   for file in {access.log,traffic_router.log}{,.1}; do
     for ((i = 0; i < 117528; i++)); do
       echo "$file" >> "$file";
     done;
   done
   sync;

3. Start Traffic Router and verify that traffic_router.log rolls over, resulting in the old traffic_router.log being moved to traffic_router.log.1 and the old traffic_router.log.1 being removed
4. Send requests to Traffic Router until access.log rolls over 10 times.
5. Verify that you end up these logs related to access.log:

   access.log
   access.log.1
   access.log.2
   access.log.3
   access.log.4
   access.log.5
   access.log.6
   access.log.7
   access.log.8
   access.log.9
   access.log.10
   

If this is a bugfix, which Traffic Control versions contained the bug?

• 6.0.2
• 5.1.5

PR submission checklist

• This PR has tests
• This PR has documentation
• This PR has a CHANGELOG.md entry
• This PR DOES NOT FIX A SERIOUS SECURITY VULNERABILITY (see the Apache Software Foundation's security guidelines for details)

[srijeet0406]

Code looks good. Tested manually on CIAB to make sure the output is as expected. LGTM!