1 in almost all cases below it is recommended to use the -j option to make
2 which was the point with uing make in the first place
3 so that maximum degree of parallelism can be reached
5 ==================== usual target
7 shortcut for 'make init all'
9 create util subdirs needed by this tool
11 initialize the whole topology on all nodes
12 ==================== inspect donfig
14 show SLICE as defined in conf.mk
19 show nodes and links declared in conf.mk in various formats
22 try out ssh connectivity to slivers in the slice, or show installed ovs version
23 make sshs SSH-COMMAND="bla bla bla"
24 run "bla bla bla" in all slivers (and under sudo)
25 ==================== For changing the topology on the fly
27 create a link between nodes with ids <id1> and <id2>
29 delete link between nodes with ids <id1> and <id2>
30 ==================== cleanup
32 tear down everything on the slivers, stops daemons and all
34 clean up whatever has been created on the slivers, keeps daemons running
36 clean up local cache, useful when restarting from a clean slice
37 ==================== configure (direct switches towards) OF controllers
39 show how OF controllers are configured from conf.mk
40 use CONTROLLER_<nodeid> if set, or CONTROLLER as a default otherwise
41 e.g. CONTROLLER=tcp:112.34.23.56:6633
44 configure the various switches so they use the configured controllers
46 reset the switches to run in standalone mode
47 ==================== devel
49 push a new version of sliver-ovs into the slivers
50 use SLIVER_OVS that should point to the local new version
51 ==================== manually/progressively deploy stuff (done in make all)
53 initialize ovs db in all slivers
55 start ovs switch in all slivers
57 create a bridge interface linked to tap device in each sliver
59 shortcut for make dbs switchs
61 stop all ovs switch in all slivers
63 list status of db and switch processes in all slivers