Packaging Harvest cached
Good day folks,
I have packaged the Harvest cached, a http, ftp, wais and gopher proxy
server. According to the documentation it is the best. After testing I
came to the conclusion that it it at least much better than the procy
from Cern.
I would like to package it for Debian as well. Before doing so I want
to hear if someone else is working on it. And if not, I would like you
to answer the following questions:
1. Where should the cache files go?
I have chosen /var/spool/proxy on my system, but you might want
them somewhere else.
2. Where should logfiles go?
I have choosen /var/log/cached/*. These are two single files so
they might also fit in /var/log as well. What do you want?
3. Rotating logfiles.
Harvest cached can rotate logfiles itself, but it does not compress
them. I have disabled this feature to rotate them manually (via
savelogs), cached can reopen the files, so that's no
problem. Should I do the same for Debian?
4. Where should the binaries/libraries go?
I have choosen /usr/local/lib/harvest which might not be a good
idea. Anyway what do you want there?
5. The files are owned by nobody.nogroup, do we have them and which
values do they have.
6. cached normally is started by a controlling script, if cached
should exit the script starts it again. This can be disabled if it
is installed correctly. Should I disable it?
7. Harvest cached is a little bit freaky. Using the original code you
cannot run it in an environment that isn't connected to the big
internet. It does some tests with the nameserver.
How should I mention this? It depends on any nameserver, it is not
useful on a single site.
The above mentioned tests will probably fail on sites that are
connected to the internet only temporarily and cause the cached to
quit.
Now, what should I do? Shall I disable the tests? Shall I change
the hostnames? (I did that for me, we have two nameservers in out
net, but no internet connection). If so, which hostnames should I
use instead?
8. Shall I enable the proxy for everyone? (allow all) Or should I deny
acces to it by default and let only hosts on the local net use it?
If so, how shall I determine the local network?
9. Shall I try to get the domains from /etc/resolv.conf and define
them as local domains? (documents from that domains are only
stored a few hours instead of some days)
10. On which port should the cached listen?
The default is 3128. I choose 81 for myself. (88 is reserved for
kerberos) 8000, 8080 and 8086 are also used by several sites.
11. On removal, shall the cache be emptied and all logfiles thrown
away?
12. The original version is 1.4.pl0, does this make problems with dpkg
versions? I haven't found something on this topic in the
Guidelines. If I don't get an answer I will assume that there are
no problems.
BTW: There are a bunch of documents concerning Harvest cached, I seem
to remember that they only exist free as postscript files. Shall I
fetch and install them, too?
I would be very pleased getting answers from you. I will start
packaging this for Debian when master is open for uploads.
Have a pleasant weekend,
Joey
PS: To avoid confusion, I don't have a real Debian system up and
running here. Instead I have a debianized system here. It's based
on a very old Smackware which I have updated manually. Some newer
packages use the Debian dpkg mechanism.
--
/ Martin Schulze * joey@oldenburg.linux.de * 26129 Oldenburg /
/ http://home.pages.de/~joey/
/ Ich glaube nur der Statistik, die ich selbst gefdlscht habe! /
Reply to: