On Mon, Sep 19, 2011 at 2:45 PM, Christer Edwards <[email protected]> wrote: > Guys - > > I spoke with Karen this afternoon and I told her I'd get her a list of > things we wanted to start getting done in the near future. I'd like > feedback from everyone on what the priority items are and the timeline > we expect. > > The first things in my mind are in regards to the new clipboard > server. This is supposed to be the new LDAP as I remember. What else > was bound for that box? >
Clipboard is the smaller Dell R310, which was bought to host ldap + dns + mango and nothing else. Once these services are completely moved over, we can retire label. If someone were to transfer mail to another physical server or (preferably) a vm, we could retire menubar as well. This would be 2 older servers gone. > Where would we like to get started? I can make some time beginning > this evening if anyone else is around and wants to dive in with me. Clutter is the beefier Dell R610 which is configured similarly to combobox with lots of ram and drive space. It would be good to see a "webapps2" vm on clutter to spread the load off of the webapps vm and to have a viable DR backup of some sort. The "original" plan was to use combobox as a replacement for container. > Let's just throw out a list of things that we need done and go from there. So perhaps a starting action item list with some random brain dumping included: * Finish the base configuration of clipboard and clutter * Move ldap and mango from label to clipboard, making sure to update the ldap slaves. * Move nfs off of container onto combobox. * Find or create a place to move mail + mailman and move them off of menubar. A vm seems ok for me, but this is a team decision. * Move all of the databases off of combobox back onto drawable. Bonus points for figuring out a sane failover strategy for when drawable fails again. * Find what else needs to be moved off of container to retire it. * Put in a formal request with Redhat IT to reap label, menubar, and container. * Set up an ldap slave in a vm and update the sssd configuration in the Redhat datacenter to point at both * Make sure all of the Ubuntu servers are fully puppetized if they aren't already * Standardize the Ubuntu servers on the most recent LTS release and set some of them up for DR from the Redhat datacenter. * Split the load between the webapps vm on combobox and the to be created webapps2 vm on clutter. Since combox will (eventually) contain the primary NFS store, lets put the more IO intensive vms on clutter if possible. It might make a lot of sense to setup a small (low memory footprint, 1 vcpu assigned) virtual machine dedicated solely to QA or the reports that fredric peters likes to run on the builds. This should be a good starting point. What do you have to add? > Thanks, > Christer > -- Jeff Schroeder Don't drink and derive, alcohol and analysis don't mix. http://www.digitalprognosis.com _______________________________________________ gnome-infrastructure mailing list [email protected] http://mail.gnome.org/mailman/listinfo/gnome-infrastructure
