sysf v1. Version 2 of sysf is compatible with version 1, but as the later version
has more features, here are some points that one must be aware of.
sysf now reads the configuration file /etc/sysf.cf.
After installing the new version of sysf one must thus create and
initialise that file as described in the installation procedures.
sysf version 2, the config name (the fourth
column in the file sysf.conf) was never used (unlike what the
documentation claimed). The entry name (1st column of sysf.conf)
was used instead. This bug has been fixed and the config name is used
as the prefix for the config file. Before using the sysf version
2 check the file sysf.conf and make sure that the config name
is identical to the entry name for every entry (one can then change
this later if required).
sysf now maintains the ownership and the access modes of the
configuration files. This means that running the check sysf -check
with sysf version 2 will display many discrepancy while version 1
displays none. One must thus check every file on the system and archive
the ownership and the access modes when the properties of the
config file differ from the system file.
sysf can be used to enforce that a file does not
exist (the config file is a symbolic link to NONE).
For this reason, sysf now expect to have a config file and a
system file for every host unlike version 1 of sysf which
assumed that the combined absence of the config and system file
was intentional.
sysf can use ssh instead of rsh to execute commands on
remote hosts.
sysf does a few extra consistency checks before performing any
command. In particular, each config groups must have a directory in the
config directory and config groups are not allowed to have the
same name as an entry name.
entry names now produces an error message when the
command modifies a config or a system file. For the test commands,
an empty list is interpreted as a the full list of entry names
(i.e. the same as ALL).