From Fedora Project Wiki
(Created page with "= Interview #1: kdetony = * Works at a big energy company * Works on a team of 6 sys admins, 2 in Peru, 4 in Chicago * Manages approximately 1000 servers ** Half and half phy...") |
mNo edit summary |
||
Line 1: | Line 1: | ||
= Interview Queue = | |||
Folks who have volunteered to be interviewed but we haven't interviewed yet: | |||
* davidstrauss | |||
* jsmith | |||
= Interview #1: kdetony = | = Interview #1: kdetony = | ||
Latest revision as of 18:39, 26 November 2013
Interview Queue
Folks who have volunteered to be interviewed but we haven't interviewed yet:
- davidstrauss
- jsmith
Interview #1: kdetony
- Works at a big energy company
- Works on a team of 6 sys admins, 2 in Peru, 4 in Chicago
- Manages approximately 1000 servers
- Half and half physical/virtual (KVM and Xen Cloud)
- All Java
- Each environment is cloned, the differences are in the versions of mod_jk and/or JDK
- Load-balancing configuration
- Changes managed via scripts (shell vs python or perl depends on the complexity) - company policy requires these are written and maintained in-house only
Goals:
- Automate all the things!
- Better understand the status of all systems, including installed software and utilization of system resources.
- When modifying a system, define a recipe that describes the changes to make, then later do a diff on the system to see the correct functionality
Motivation:
- Likes the idea of being able to customize on the needs one can have
Frustrations:
- Production pushes where the JDK crashes.
- Production pushes where the VM has to be taken down and have its capacity raised before the new version of the application can be started.