Interesting how sysadmins/coders get production systems, especially on LAMP and HA cluster setups online without a Performance/Scalability Test - case in point, you have a very good application (lots of $$ expected once its launched) and mounting pressure from the senior management to 'get things working!!' A couple of minutes, hours, days, weeks or months after launching the system(depending on the amount of load you're dealing with) and you realize: 1. Did I need a 3-way failover plan? DRBD only works with 2 nodes. Should I have slaves replicating off the replicated databases? 2. Mysql Replication or DRBD / DRBD+Heartbeat.or DRBD+Mysql Cluster - how often Im I making schema changes on my application(s)? 3. For my Apache setup maybe FastCGI was better to use than mod_php? 4. Ok.. its a network issue, do I have a technical solution to improve resilience and network load? 5. IM I LEARNING FROM A PRODUCTION SYSTEM? *Something I observed as I took backstage while constantly giggling and later left in my capacity as volunteer-until-the-system-runs-then-we-can-discuss-employment-terms in an upcoming IT firm. Until I get a serious employer: $cd /pub $more beer $find /share/.trash/ -type f -atime +30 -exec rm -fr {} \;