(Imported from MoinMoin)
m (1 revision(s))
Revision as of 16:37, 24 May 2008
#!html <div style="height:66px; width:100%; background-color:#002867;"> <a href = "http://fedoraproject.org/wiki/Infrastructure"> <img style="float:right;padding-top:3px;" src="http://fedoraproject.org/wiki/Infrastructure?action=AttachFile&do=get&target=InfrastructureTeamN1.png" /></a> </div> <HR style="height:2px; background-color:#00578E;" />
What is a FIG?
A FIG (Fedora Infrastructure Group) is a cluster of people that work together on a specific part of Fedora's Infrastructure. The organizational structure of a FIG is simple. Sponsors, non sponsors. All of whom have equal access to the environment with two exceptions. 1) Sponsors can create new files on the machines (instead of just alter) and 2) Sponsors can grant access to other people for whom they are responsible.
Users in each FIG have access to the machines themselves as well as access to the CVS that hosts configuration management. Each FIG is part of the Fedora Infrastructure Project and all are part of the same team.
What are the FIGs?
The sysadmin build group is responsible for the entire build environment. This includes the plague and koji masters as well as the various build machines (currently ppc and x86_64). This group has shell access to the machines and configuration environment.
All members in the sysadmin-cvs group exist to maintain our source control environment and related machines (presently only the cvs box). This includes all packages and lookaside cache.
All of our web sites are maintained by the sysadmin-web group. This includes all web site application servers as well as the proxy servers. This is our busiest group as most of our active changes is on web servers, which comprise the largest number of servers.
The sysadmin hosted group maintains the servers that run Fedora Hosted projects . Team members handle the creation/administration of Hosted projects/repos and related services.
The Release Engineering team has a set of servers it uses to create Fedora. This group has shell access to these servers and related configs in puppet.
The sysadmin-main group is the group that has root access to all other machines in our environment. This group is closely monitored and access is highly restricted, no one is admitted directly into the sysadmin-main group without first proving themselves in another group.
The DBA group has access to all of our database and LDAP machines. Due to the highly sensitive nature of the database boxes this group, like sysadmin-main, is highly restricted.
sysadmin-test grants access to our bed of test-servers. The test servers (like publictest1) are primarily used to do proof of concept implementations for Fedora Infrastructure or other projects. Anyone filling out an RFR needs to get into this group in order to prove their system should be used in Fedora's production environment.
A great deal is done with development in Fedora's Infrastructure. This group, like sysadmin-test, has access to all of our test servers. It also grants write access to the sysadmin repo. Working on our scripts and code is especially important when getting to know the group.
The noc team has access to nagios and cacti (and whatever other monitoring tools we are using) They also have read only access to many of the systems to poke around and see what might be wrong as well as find things that need to be monitored.
Part of the Infrastructure teams job is to provide tools to make our developers and contributors more efficient. The sysadmin tools group is in charge of these services. Such services might include asterisk (VoIP), gobby, mailman, etc. This team works closely with the sysadmin-web team.
The sysadmin group is the base group required by all other FIGs. Nagios notifications are sent to this group, it does not provide any additional access but is a good way to welcome yourself to the team.