That, at least, is what it looks like to me.

I have the BOINC client installed on my FreeBSD box. I can attach to projects just fine, but when I try to start any of them, I get the following messages (or something similar, depending on the project):

11/30/2005 1:39:44 AM||request_reschedule_cpus: project op
11/30/2005 1:39:47 AM|rosetta@home|Sending scheduler request to http://boinc.bakerlab.org/rosetta_cgi/cgi
11/30/2005 1:39:47 AM|rosetta@home|Requesting 8640 seconds of work, returning 0 results
11/30/2005 1:39:48 AM|rosetta@home|Scheduler request to http://boinc.bakerlab.org/rosetta_cgi/cgi succeeded
11/30/2005 1:39:48 AM|rosetta@home|Message from server: platform 'i386-portbld-freebsd5.4' not found
11/30/2005 1:39:48 AM|rosetta@home|No work from project
11/30/2005 1:39:48 AM|rosetta@home|Deferring communication with project for 23 hours, 59 minutes, and 59 seconds


It appears that the server refuses to send work units to the client because it doesn't recognize the O/S on which it runs. (Message from server: platform 'i386-portbld-freebsd5.4' not found). Short of recompiling my kernel in order to change the reported O/S string, is there anything I can do to fix this?

---

You are close but missing one little point. The servers recognize your OS however there is no science application for that OS. As far as I know none of the projects have a native BSD app. You can compile your own version of the seti application and use the anonymous platform but none of the other projects have released the source code.

---

BOINC WIKI

BOINCing since 2002/12/8

As far as I know none of the projects have a native BSD app. You can compile your own version of the seti application...

I compiled my BOINC client from the FreeBSD ports collection, so perhaps it's simply an oversight when it hasn't been set up with the anonymous platform identifier (which did resolve the issue), "out of the box." However, since the BOINC client (including the setiathome science application) has been incorporated into the FreeBSD ports collection, and as such is now "officially" a part of the native software base, perhaps it would be a good idea to add a platform identification expression such as for example /i386-\\w+-freebsd-\\d\\.\\d+/ to at least the setiathome project?

---

I'm confused now - you apparently have it working, but do you have it working for Rosetta? Or just SETI? I had believed that only SETI could be run on FreeBSD...

---

I'm sorry about the confusion. It dawned on me previously that rosetta was not an option, so I only have it working for setiathome, obviously, although that wasn't entirely evident from my reply.

What I am concerned about now is the fact that my FreeBSD box impersonates a Microsoft Windows machine toward setiathome. I haven't been able to find anywhere in the Wiki confirming that work units transmitted to the science apps are platform independent, but if this is really going to work in the long run, I guess they must be. Do you know?

---

I haven't been able to find anywhere in the Wiki confirming that work units transmitted to the science apps are platform independent

It depends on the project; some send out different work to different platforms, which can either be identical or different. In the case of SETI, the same WU is sent to "whoever asks for work next", so they are straight text files, platform independent.

As for adding FreeBSD to the SETI "known" list, I'm not sure they feel it's "stable" enough to do that, unless they were providing the software direct from UCB. I don't think any of the "non major" platforms are on the list, I think HP/UX, Solaris, etc., all use the same mechanism as FreeBSD.

---

As for adding FreeBSD to the SETI "known" list, I'm not sure they feel it's "stable" enough to do that...

Now that you mention it, that's apparently a good point, since my FreeBSD client consistently croaks after downloading WUs with a "setpriority: No such process" message. Maybe I just need to wait for the next release of BOINC for FreeBSD.

---