Installation instructions for rebatch v1.0
Read the file README for a general
overview of rebatch. The file
THEORY explains how this program hangs
together at the technical level.
Contents
The rebatch package requires
Perl and its libraries to be
installed on your system. It runs under both Perl 4.036 and Perl
5.000 without modification. It does not use any Perl 5 specific
features, or hairy Perl constructs ;)
All of the configuration for rebatch and
where.to is set in Perl variables in the file
rebatch.conf
There are also several variables related to the installation of
scripts and manual pages in the Makefile.
The steps below outline what needs to be looked at and what it all
means.
READ THE ASSUMPTIONS
PART OF THE THEORY FILE. If those
assumptions are not true at your site you are going to have to work a
bit harder. I give some work arounds. If you find any more, please
let me know.
While you are at it you might as well read the rest of this
installation document, the README and the manual pages.
The rebatch package can be auto-configured using the
values set in INN's config.data
file. To
make use of this:
(a) Edit the Makefile
and change the definitions at the
top of the file to point to the INN files at your site:
- INN_CONF_DIR
- the 'config' directory in INN's distribution
- INN_SUBST
- location of subst program
- INN_CONFIG_DATA
- location of config.data file
- SED
- same as in config/Makefile
(b) Type
make config
to perform the substitutions.
You should definitely look through rebatch.conf
and
the Makefile
to make sure it has done the right thing,
and the values there match the reality of your setup.
(a) To perform automatic installation using the Makefile you will need to
alter the variable
INN_INSTALLIT
to point to the installit.sh script in INN.
If you are unhappy with the installation directory you will need to
alter the line:
NEWSBIN = /usr/local/news/bin
to point to the right place.
The first line of the rebatch and where.to scripts
#!/usr/local/bin/perl5 -w
may need to be altered to point to your Perl interpreter.
The '-w' flag makes Perl complain about possible typos or
logic errors, as well as just plain stupid mistakes I have
made. Remove it if you wish.
If you altered the NEWSBIN
variable in the
Makefile
you will need to adjust the lines in
rebatch and where.to that read:
push(@INC, "/usr/local/news/bin");
to match the Makefile
5. All of the configuration parameters
All of the variables used by the rebatch and
where.to scripts are set in the file
rebatch.conf
You should look through these and make sure they match your set up.
The variables are:
- $debug
- debugging level. Also set by the
-d
and -n
flags
- $remove_status
- should rebatch remove its temporary files when the
batch has been sent? See the notes on
$rebatch_status_prefix
below.
- $path_batchdir
- where does INN write batch files
- $path_newsfeeds
- where is your newsfeeds file?
- $path_locks
- where to write lock files.
- $rebatch_lockfile
- lockfile for the master rebatch process
- $rebatch_lockfile_prefix
- lockfile prefix for each site behind. Lock files are make of the
form
$rebatch_lockfile_prefix.$sitename
, e.g.,
/usr/local/news/LOCK.rebatch.auckland.ac.nz
(for $sitename == auckland.ac.nz).
- $nntplink_id
- how to identify NNTPlink feed sites in newsfeeds file
- $nntplink_link_suffix
- NNTPlink state files that appear in the
$path_batchdir
directory. These are the files with the
name like news.auckland.ac.nz.link
.
- $nntplink_tmp_suffix
- suffix NNTPlink uses when remote host is down. These are files
with the name like
news.auckland.ac.nz.1234.tmp
.
- $log
- log file for rebatch. Set to
"
/dev/null
" if you don't want any. You might want to add
rotation of this file to your news.daily progress.
- $rebatch_batch_suffix
- suffix used for rebatch batch files in
$path_batchdir
.
- $rebatch_status_prefix
- prefixed uses to record innxmit's debugging output.
These files can get big and DO NOT GET DELETED when the innxmit
finishes (why? so you can easily see the reason the site stopped)
Choose somewhere where there is lots of space. You can make
rebatch delete them by default or by a command line
option.
- $ctlinnd
- where do I find this program?
- $innxmit
- how to send off batch files. It is probably best to include a
'
-t nnn
' timeout parameter to avoid hung processes.
where.to has a subset of these variables.
Aftering altering the variables, you can type
make install
to install the scripts in the directory given in the Makefile.
You can run the programs in the current directory by using
perl rebatch parameters
or
chmod a+x rebatch where.to
./rebatch parameters
7. Test it out
Try rebatch out in 'debug, no execute mode' to see what it does...
./rebatch -n | more
it should print things out about all the sites it finds, then what it
thinks about each site. It will tell you about the files it finds and
what it will do.
If it pauses after a saying
.... running innxmit -d nntp.host nntp.host.rebatch
don't panic. There is a small sleep in there to make the
parent/child waiting stuff works.
You could then try it in real mode
./rebatch -d
and then try tailing /tmp/rebatch.whatever
to see it in
action. Then try out 'where.to'
8. Install the manual pages (optional)
To install the manual pages, type
make install-man
De-Installation
If you ever want to uninstall this package you can type
make uninstall uninstall-man
or remove the files
- rebatch
- where.to
- rebatch.conf
- rebatch.common
from the binary directory and the man pages
from the manual page directory.
Part of the rebatch package
Russell Street
(r.street@auckland.ac.nz)
Last updated: 12th February 1995