Furosh One wrote:
Furosh,
See my relies in line.
Okay so I've spent a few days and nights going through and learning
fossology as much as possible. Trust me I've spent all last weekend
going through docs, newsgroups, mailing list archives... and can't
find best practice for my scenario =/
I believe I'm on the right track, this is what I got so far.
Cluster-
- (6) PCs running Ubuntu v9.10 - installed Fossology with 'sudo
apt-get install fossology'
- Host file setup to use ssh to host1, 2, 3, 4, 5, 6 and I tested SSH
auto-login.
- Currently have DB on each machine with Apache web server (since
apt-get install).
Question anyone can answer:
1.) I think I would have to copy the Hosts.conf file to every PC as
distributed repo model. Correct?
(host 1 * 00 2a, host2 * 2b 54, host3 * ff 7f, etc, etc,)
Yes, let's talk terminology. You will have a db server, a web server
(both can be on the same machine). Then you have agents.
The scheduler runs on the db machine. The scheduler uses ssh to talk to
the agents.
All machines must have the same *.conf files EXCEPT Scheduler.conf. The
agents do not need that file.
so you will see this in ...etc/fossology on the agent machines:
Db.conf Depth.conf Hosts.conf Proxy.conf RepPath.conf
2.) I was thinking I need to have one central DB in this config so I
figured I could shut down the postgres DB on each agent and apache as
well, but my machine I want DB/ApacheUI has FATAL errors. Do I need to
keep the DB running on each PC?
What are the errors? Are they due to fossology or something else...
NO! in fact, you should not do so. The agents only need the postgres
client software. They don't need apache either.
3.) If so, I should modify Db.conf to all point to DB/Apache server
"host=host5;" ?
4.) Do I put this on all machine's DB.CONF file?
The Db.conf files should all be the same and should all point to ONE
machine that is acting as the db server.
5.) Lastly, how would I configure the Scheduler.conf for this, just
One or on ALL PCs (I would think just one)???
Yes, you are getting it. Just one. The one where you are running the db.
Currently my Scheduler.conf has a line like this for EACH host/agent
(host1, host2, host3, host4, host5, host6):
Below seems almost right... host1 shoudl be clear6, the hosts.conf file
uses the hostnames (without the domain stuff e.g foobar, instead of
foobar.baz.com) of each agent.
In fact, on the agent machines you can just install:
apt-get install fossology-agents fossology-common
That's all they need.
Another important part is that all the agent machines repo area must be
nfs mounted to the server where the db/repo is. AND all the agents MUST
cross mount all the other agents repo storage areas.
The repo is at: /srv/fossology/repository/<machine-name> or localhost
(localhost is used for single system set ups, you can ignore).
Here is typical muti-config we use:
Server BOSS: runs the scheuduler, db and web server
Agent: wkr1: runs the agents, has a part of the storage area (that's
what the Hosts.conf file is splitting up).
Agnet: wkr2: same as agent wkr1 (see below)
The server has storage areas for the repo:
/srv/fossology/repository/wkr1 (mounted from wkr1)
/srv/fossology/repository/wkr2 (mounted from wkr2)
Agent wkr1 has storage areas:
/srv/fossology/repository/wkr1 (local disk)
/srv/fossology/repository/wkr2 (mounted from wkr2).
Agent wkr2 has storage areas:
/srv/fossology/repository/wkr1 (mounted from wkr1)
/srv/fossology/repository/wkr2 (local disk)
The Hosts.conf looks like:
wkr1 * 00 7f
wkr2 * 80 ff
Setting up a multi-system like this is not for the faint of heart! It
takes a lot of system admin knowledge to be able to do it. It's a lot
of work for us as well. But it does allow you to put significant
computing resources into the analysis.
Keep those questions coming, we will get you set up. Hope this helps.
=========================================
%Host host1 2 1
agent=wget host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/wget_agent -d /var/lib/fossology/agents"
agent=unpack host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/engine-shell unpack
'/usr/lib/fossology/agents/ununpack -d
/srv/fossology/repository/ununpack/%{U} -qRCQx'"
agent=adj2nest host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/adj2nest"
agent=filter_license host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/Filter_License"
agent=filter_license host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/Filter_License"
agent=license host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/bsam-engine -L 20 -A 0 -B 60 -G 15 -M 10 -E
-T license -O n -- - /var/lib/fossology/agents/License.bsam"
agent=licinspect host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/licinspect"
agent=mimetype host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/mimetype"
agent=mimetype host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/mimetype"
agent=specagent host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/specagent"
agent=filter_clean host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/filter_clean -s"
agent=delagent host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/delagent -s"
agent=sqlagent host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/sqlagent"
agent=sqlagenthost host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/sqlagent -a sql"
agent=pkgmetagetta host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/pkgmetagetta"
agent=pkgmetagetta host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/pkgmetagetta"
agent=fosscp_agent host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/engine-shell fosscp_agent '/usr/bin/cp2foss
%{*}'"
agent=fo_notify host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/engine-shell fo_notify '/usr/bin/fo_notify
%{*}'"
agent=selftest host=host1 | /usr/bin/ssh fo...@clear6
"/usr/lib/fossology/agents/selftest -s"
============================================
Matt Taggart sent me a good link to get me started - Thanks!, now I'm
a bit stuck to complete this...
-Regards,
FuRoSh...
------------------------------------------------------------------------
_______________________________________________
fossology mailing list
[email protected]
http://fossology.org/mailman/listinfo/fossology
--
Mark Donohoe
OS&T, Cupertino CA.
fossology.org
_______________________________________________
fossology mailing list
[email protected]
http://fossology.org/mailman/listinfo/fossology
