Start the first node's server process

[root@octopus ~]# systemctl restart stasher.service
[root@octopus ~]# links --dump http://localhost/portmap
   service,user,pid,prog,port
   httportmap.libx,0,26741,/usr/sbin/httportmapd,/var/httportmap.client
   octopus.objrepo.example.com,2,10420,,49329
   pid2exe.libx,0,26741,/usr/sbin/httportmapd,*
[root@octopus ~]# stasher
Ready, EOF to exit.
> admin
Connected to objrepo.example.com, node octopus.objrepo.example.com.
Maximum 10 objects, 32 Mb aggregate object size, per transaction.
Maximum 10 concurrent subscriptions.
octopus> status
Cluster:
  Status: master: octopus.objrepo.example.com (uuid S85HVpjTVrQqA000_F1aJm00001vhGO00318n4AS), 0 slaves, timestamp 2012-03-17 16:15:56 -0400

... long output deleted ...

octopus> [EOF, CTRL-D]
[root@octopus ~]#

The default stasher package installs a systemd service that starts a stasher daemon process for every node directory in /var/stasher/nodes when the machine starts. After adding the first node directory to /var/stasher/nodes, systemctl restart restarts the service, which starts stasher in the newly-installed directory.

If not using the default stasher package, for some reason, the stasher.sysinit script in the source tarball shows the recommended steps for starting and stopping stasher:

A succesful stasher starts is confirmed by verifying that the node is listed as an httportmap service (octopus.objrepo.example.com in the above example), and by connecting to the server, as in the shown example. Open http://hostname/portmap in a browser, and verify that the service is listed.

The admin stasher command takes the pathname to the cluster node directory as a parameter. If there's only one subdirectory in the default stasher node directory (/var/stasher/nodes or /usr/local/var/stasher/nodes, depending on stasher's configuration), the directory parameter may be omitted, as in this example.