4 Replies Latest reply on Jan 25, 2013 10:02 AM by Joe Piotrowski

    BladeLogic 8.2 Configuration on Linux 64-Bit

      We are currently in the processes of upgrading out infrastructure that BBSA sits on. We are running BladeLogic 8.2 SP1 on RH Linux 5. On our old infrastructure, our job servers are split into 2 instances on one physical server. Each has a max heap size of 4GB.


      Is there any advantage to keeping this type of configuration, 2 job server instances on one physical server, or will I get the same performance if I defined 1 job server instance with an 8GB max heap size (also increasing the max work item threads, etc...)?


      I would much prefer to configure each server as one large job server. This will make remediation of a server going down much more practical since there is no command to restart only one logical BL app server when you have multiple instances defined and running on one physical server (no command I have found yet, anyway). Currenlty, I have to restart the entire server (/etc/init.d/blappserver stop/start) to restore one instance. This carries the obvious risk of losing jobs running on any other instances on the same physical server. We have a script that monitors each server instance and sends an SNMP trap to Atrium Orchestrator when an instance goes down. Right now, AO waits until all the instances on one server go down before it initiates a restart of the server. If I only had one large instance on a server, AO could act immediately.