Yes Lars makes good warning about increasing max_connections means you need to take into account increased memory use per connection. Which is mostly affected by the work_mem setting.
A few things to consider : 1. perhaps you only need to run a training server when training. 2. you might not need to run a training server per instance 3. if you are using a staging server (eg for doing metadata editing) then you probably are not going to have many concurrent users for this so you can tune down all pool settings. Lars, I haven't looked very recently, but to what extent are we creating new jdbc connections outside of the c3p0 pool? If we are only strictly using the pool its easier to provision these things. On 9 April 2014 14:23, Lars Helge Øverland <[email protected]> wrote: > > Okay. So each DHIS instance will require its own set of connections, so > you should take required connections per instance and multiply with number > of dhis instances for your max_connections. > > Running 12 dhis instances on a single database servers sounds a bit too > much, then the db must have enough mem to support close to 1000 > connections. So bumping your max_connections will fix the problem, but > might lead to trouble down the line. > > max_connections = 900 > > _______________________________________________ > Mailing list: https://launchpad.net/~dhis2-users > Post to : [email protected] > Unsubscribe : https://launchpad.net/~dhis2-users > More help : https://help.launchpad.net/ListHelp > >
_______________________________________________ Mailing list: https://launchpad.net/~dhis2-users Post to : [email protected] Unsubscribe : https://launchpad.net/~dhis2-users More help : https://help.launchpad.net/ListHelp
