Acumen indexing is automated by a cron job scheduled under the "acumen" username, which runs a script in /home/acumen/bin called acumen.pl
This script does the following things:
- Backs up the primary acumen database
- Switches out the index.php scripts in the primary web directory (/srv/www/htdocs/acumen-old/ ) so that the staging subdirectory becomes what is live online instead, and the staging database is the one referenced for delivery
- Performs a solr index on the acumen database (java -Xms1g -Xmx2g -jar /home/acumen/acumen.jar acumen >> /home/acumen/acumen.log) – (NOTE: I recently raised the amount of RAM allotted in this call, so you shouldn’t have to modify it again unless the quantity of content grows a great deal before you switch systems.)
- Dumps copies of the primary tables out of the newly indexed database
- Re-switches out the index.php scripts so that now the acumen-old directory is again what is live online
- Pulls copies of the primary tables into the staging database
- Imports that data into the staging version of solr
- And removes older versions of the backed up databases and tables. It also logs what it’s doing.
It’s been working fine without modifications for almost a year now, and hasn’t needed any attention.
What’s more likely to go wrong is that content will be mucked up that gets put into the web directories – and that has to be corrected. Sometimes that means correcting what’s in the database as well as the web directories, and then reindexing.