ndmfs is an implementation of the proposed File Server extension to NDMP version 4.
launch ndmsrv on start (listen on port 10000 by default) - static binary in /opt/NDMP
To run the server type './ndmsrvr'. You can optionally specify -d
for debug output, or -p to specify the tcp port the server
should listen at (default is 10000). You must run the server as root.
The Linux host must be configured to use an /etc/shadow password
file if you wish to use TEXT authentication. The effective uid and
gid of a server session is set to 'nobody' for auth_type NONE,
and the specified username for auth_type TEXT or MD5, thus for complete
functionality you should create the specified username account on
the Linux host. The server "exports" (via NDMP_CONFIG_GET_FS_INFO) all
filesystems mounted on the Linux host (whatever appears in /etc/mtab).
The MD5 shared password is hard-coded to abc123.
To run the DMA, type './ndma'. It provides a shell and should prompt
you with 'ndma> '. Type 'h' or '?' for help, or '? all' for complete
descriptions of all supported commands. With ndma it is possible to
establish 3-way NDMP sessions and perform backup and recover operations
between NDMP File/Data Servers, as well as exercise all NDMP File Server
RPC's. For full functionality ndma should be run as root (SERVER_AUTH
of auth_type TEXT requires /etc/shadow access).
To run the expect script, edit the config file 'ndma_exp.conf' as needed
for your setup (You will need to make sure that the files referred to
by the variables EXISTING_RW_FILE and EXISTING_RO_FILE are created and
can be accessed by the 'ndmp' user, and that the file referred to by
the variable NON_EXISTING_FILE does not exist. Probably nothing else
needs to change). To start the script, as root type './ndma.exp all
stop_on_error 1'. This will run a battery of test cases against the
server 1 time and stop if any error is encountered. The last argument
specifies the number of iterations to run. You can substitute 'all'
with 'connect', 'config', or 'file' to exercise just those areas of
the server. Anything other than 'stop_on_error' will cause the script
to continue running when an error is encountered.
When problems arise check your syslog file. Both ndmsrvr and ndma log
error conditions with syslog (moreso with ndmsrvr, as ndma usually just
prints things to stdout).