proofd - Online in the Cloud

PROGRAM:

NAME

proofd - PROOF (The Parallel ROOT Facility)

DESCRIPTION

Using PROOF (The Parallel ROOT Facility) one can analyze trees in parallel on a cluster of
computers. The PROOF system consists of the proofd(1) from-end program which is started
via inetd(8).

proofd takes care of user authentication and overlays itself then with the desired version
of the proofserv(1) executable.

The proofserv(1) is a basically the same as the ROOT interactive module root(1), except
that it reads its commands from a socket instead of from the terminal.

Since it is a remote server it will not do graphics and therefore is not linked with any
graphics libraries (Motif, X11, etc.).

SETTING UP PROOF

Since you need to handle sensitive system files, you need to have root (the user, not the
program!) privileges. Also, it's recommended that you read the appropriate man(1) pages
first. These are inetd(8) (the internet daemon), sysklogd(8) and syslog(3) (the system
logger daemon), and init(8) (the SYSV process control initializer).

In other words:
Be very VERY carefull when installing proofd
There, I said it. And ofcourse, no warrenties what so ever.

1 Make sure every node has ROOT installed. We'll assume you installed the ROOT
applications in <bindir> (e.g. /usr/bin) and the PROOF configuration files in files
in <proofdir>/etc (e.g. /usr/share/root/etc).

2 The TCP port 1094 was allocated by IANA, ( www.iana.org ⟨Iwww.iana.org⟩ ), to
rootd, so add to /etc/services the lines:

proofd 1093/tcp
rootd 1094/tcp

3 On each node, add to /etc/inetd.conf the lines:

proofd stream tcp nowait root <bindir>/proofd proofd <proofdir>/proof
rootd stream tcp nowait root <bindir>/rootd rootd -i

You can substitute the <proofdir> with any directory that holds your PROOF
configuration, for example /etc/root/proof. However, it should be shared among all
nodes in the cluster.

4 Restart inetd or force it to re-read the config file:

kill -1 <inetd pid>

If you use SYSV init(8) scripts, you can probaly just do

/etc/init.d/inetd restart

or whatever is appropriate for your system.

5 On the master node, add to /etc/syslog.conf the line:

local5,local6.debug <proofdir>/log/proof.log

and all slave nodes:

local5,local6.debug @<master hostname>

where <master hostname> is domain name of the master node. All PROOF syslog
messages will be collected on the master node. Just make one node in the cluster
the master, all others are slaves.

6 On all nodes, change the below lines in /etc/syslog.conf,

*.info;mail.none;news.none;authpriv.none /var/log/messages

to:

*.info;local5,local6,mail.none;news.none;authpriv.none /var/log/messages

7 Create an empty <proofdir>/log/proof.log:

echo "" > <proofdir>/log/proof.log

8 Restart syslogd or force it to re-read the config file:

kill -1 <syslogd pid>.

If you use SYSV init(8) scripts, you can probaly just do

/etc/init.d/sysklogd restart

or whatever is appropriate for your system.

9 Edit <proofdir>/etc/proof.conf to reflect your cluster configuration. See the
example proof.conf file for more information.

If you installed ROOT using some precompiled package (for example a Redhat Linux or a
Debian GNU/Linux package), steps 2 to 4 may already be done for you.

If you're running Linux, Steps 5 to 7 can be somewhat automated using the script proof-
facility in <proofdir> of your distribution. This script uses syslog-facility(1), from the
Linux syslog(1) distribution, and may run on other platforms as well - but no warrenties
mind you!

Step 9 is completely up to the user.

EXAMPLE

That's it. To test PROOF try the following:

root [1] gROOT->Proof("<master hostname>")
root [2] gPROOF->Print()
<shows information on the master and all active slave servers>
root [3] .q

Use proofd online using onworks.net services