The steps explained for basic setup enable any computer to be a functional BOINC crunching machine very quickly. However, there are still some things that have to be manually adjusted with this very basic configuration.
The more advanced configuration explained on this page will enable you to run your BOINCpe machines without any user interaction at all (beyond pressing the power button of the computer). You will be able to operate your machines diskless (without hard disk drive) and headless (without mouse, keyboard and monitor attached).
Configure BOINC Remote Access
Since you most probably want to monitor and configure the BOINC client running on the crunching machine you will have to configure remote access to this machine. The BOINC client already comes with RPC (Remote Procedure Call) capabilities. Essentially, you can use the BOINC manager GUI running on a different machine in order to connect to any other remote host (this case: BOINCpe).
The following steps only briefly describe how to enable Remote Access. Please have a look at the article „How To Control BOINC On Remote Computers“ from the BOINC Wiki. There are two text files necessary to configure remote access.
Create file remote_hosts.cfg in .\plugin\BOINC_client\data
You might have to create the file remote_hosts.cfg if you do not already have one. Actually, this file is a list of DNS names or IP addresses. Each one in a separate line. No separating characters are necessary except for the line breaks. Only the computers specified in this file will be allowed to access BOINC remotely.
See: Remote Hosts File
Edit/Create file gui_rpc_auth.cfg in .\plugin\BOINC_client\data
The file gui_rpc_auth.cfg configures the password necessary for accessing the client. It is a plain text file containing only a single line which is the password. This file is already available in directory .\plugin\BOINC_client\data. You might want to change the initial password, but you can also leave it as is.
See: BOINC Daemon Password
After putting these two files in the directory .\plugin\BOINC_client\data and building BOINCpe, your machine will be accessable by remote. If you are running multiple crunching machines in a farm, the free tool BoincView can support you in monitoring and administrating multiple BOINC machines at once…
Automatically get settings from a BOINC Account Manager (BAM)
A BOINC Account Manager (e.g. BOINCstats BAM!) is a great thing to manage your BOINC settings centrally across a number of different projects you might be attached to. You can configure BOINCpe to use your BAM account in order to start computation immediately after startup without further configuration. BOINC will then use your default BAM preferences.
You need access to the installation directory of BOINC that is already successfully connected to your BAM Account. From there you have to get the following files:
Copy file acct_mgr_url.xml to .\plugin\BOINC_client\data
This file identifies the account manager to be used (e.g. BOINCstats BAM!). Just take it from your BOINC installation directory and copy it to .\plugin\BOINC_client\data.
Copy/Edit file acct_mgr_login.xml to .\plugin\BOINC_client\data
The file acct_mgr_login.xml has two different purposes: First, it stores your personal login data for the BAM account (i.e. login name and hashed password). And then, it stores the ID of the host that connected to BAM the last time using these credentials.
After copying the file to .\plugin\BOINC_client\data we have to modify the file in order to get a new unique host ID (CPID) for the BOINCpe machine and to let it connect to BAM immediately. This is easy: Simply edit the contents of the XML tags „<previous_host_cpid>“ and „<next_rpc_time>“ to „0“. It has to look like the following:
Please remember that the host ID will be gone (this will most likely also lead to multiple machines in your project profile(s)), once the BOINCpe machine is restarted unless you configure the BOINCpe Persistency Service (see below).
Automatically backup your data using BOINCpe Persistency Service
What does BOINCpe Persistency Service do?
Are you worried about using a RAM disk for your long running computations? Want to make sure your projects never have duplicate entries for your machines again? Then BOINC Persistency Service is right for you!
It will periodically backup your BOINC data from the RAM disk to a network share on a different computer (e.g. your file server). Moreover, it will automatically restore the BOINC data from the share after bootup to start computation with the backupped settings, WUs and checkpoints. Please check the diagram on the left for further information on how the persistency service works. See the BOINCpe FAQ for further configuration of the cron scheduler, if you like.
The configuration of the BOINCpe persistency service will work fine with as many machines as you like. Simply specify the same network share for all BOINCpe machines. It will automatically create/read a subdirectory according to a MAC address of the computer.
So, do your configuration for BOINCpe once – burn it as often as you need it…
Edit your share and login data in .\plugin\BOINC_persistency\files\network_share_data.bat
You need to enter your network share and credentials for accessing it. Please follow the instructions in .\plugin\BOINC_persistency\files\network_share_data.bat.
Note: Your BOINCpe system’s MAC address will be added as a directory to path you specify in this file. This is necessary in order to distinguish between the BOINC data of different hosts.
An Example: Let us assume your first BOINCpe host has the MAC address 00-0A-0B-0C-0D-0E and your second host has the address 00-01-02-03-04-05. You can use the same path for both hosts (e.g.
\\myServer\myShare in network_share_data.bat). The final path will automatically be set to:
Also have a look at the above diagram for further illustration…
Enable the plugin „BOINCpe Persistency Service“ in the plugin menu of PEbuilder.exe
The plugin is disabled by default. You will have to enable it manually before building BOINCpe.
Note: The local data put in folder .\plugin\BOINC_client\data) will be overridden if using persistency service and valid BOINC data exists on the share. On startup, the data from the share is loaded in order to start BOINC.
You can manually load or store BOINC data using Go/Programs/BOINC/Load BOINC Data from Share… respectively Go/Programs/BOINC/Persist BOINC Data to Share…
Note: Always shut down the machine using Go/Shut down/Shut down. All BOINC data will be persisted before shutting down the computer.