As Zimbra has grown we’ve begun to see larger and larger Zimbra installs. We now have many installs with thousands, hundreds of thousands and larger running in complex multi-server deployments. Service providers in particular are driving some of the larger multi-domain and super hosting of Zimbra. Over the past couple years our engineering and support teams have learned alot.
Candace has compiled a large list of tuning parameters for our larger deployments based on field knowledge and customer feedback. Disk layout guidelines, memory allocation, OS and network stack tuning etc. She recently posted it on the wiki for easy access. It will be a living document like the rest of our knowledge-tree. While this doc mentions it’s for Zimbra installs with more than 2000 users I’m sure there are some of these settings will help those with smaller installs.
Performance Tuning Guidelines for Large Deployments
As always comments and suggestions can be posted on our Zimbra Forums.
