amarts / glfs-health Goto Github PK

glfs-health.sh - Tool to inspect the health of a GlusterFS server
==============

Usage:
   sh# glfs-health.sh [SERVERIP [SERVERPORT [TRANSPORT [VOLNAME]]]]

The tool inspects the health of the GlusterFS server volume VOLNAME running on
the host SERVERIP on port SERVERPORT using the transport mechanism TRANSPORT.

The default four parameters are initialized in the script. They can be overriden
from the command line as shown above.

The health of the service is written to stdout. The possible output messages:

-- "OK"
   The server is healthy and capable of handling client connections.

-- "Connection refused"
   The machine is reachable, but GlusterFS is not running on the requested port.
   Possible reasons include GlusterFS not yet being started, or GlusterFS server
   has crashed.

-- "Host unreachable"
   The machine is unreachable. Very likely to be shutdown, or network interface
   has gone down on the server.

-- "Server Unresponsive"
   The machine is reachable. Connection could be established at the transport
   level but GlusterFS server is not responding. This could happen for at least
   two reasons.
    1) The OS has softlocked certain subsystems but networking is operational
    2) GlusterFS has deadlocked itself due to a bug

-- "Unknown service encountered"
   The service running on the requested HOST and PORT is not a GlusterFS service

-- "Authentication failed"
-- remote-subvolume "adsf" is not found
   Server is operational as such, but the might be misconfigured or the tool is
   invoked as non-root or with wrong parameters

-- "No version number specified"
   Server is a an old version and mismatching with the client