A little while ago, about three hours now, I woke up
and, while I was up, decided to check on the SETI@home
running on my Linux box.

I clicked on the task tab ... and nothing happened.
I clicked on all the other tabs ... nothing happened.
I contemplated this for a moment, thanked my lucky
stars that I had BOINC running as a separate user,
su'ed to root from my own account window and killed
BOINC and the SETI@home task outright (kill -9 ...).

Entertainingly, BOINC or SETI@home made a face as it
died! i.e. :-^) (or something close to that...)

I then examined the BOINC directory. The slot/0
directory had a new SETI@home image in it and a new
.so file. I deleted these, opened a window for BOINC,
and started BOINC (and therefore SETI@home) again.

It is working fine with no strange behavior. Has
anyone else experienced something like this?

That's a weird nightmare you had there Christopher.

What's Linux?

I'm only joking. I see it's your first post. Welcome to the forum. I know about 'su -' and the fun you can have in the root directory, but I never saw any faces. However, I must try a kill -9 next time I go near that Linux box...

---

flaming balloons

Christopher C. Woodbury, that was the most entertaining story that i have read in a long time. I normally would not even reply to a message like this.

Chris, it was a gripping story from start to finish, i was on the edge of my seat. You have a natural talent for telling..... short stories!

I have never heard of this problem before. I think you deleted something from one of the directories and some small part of BOINC is not going to work properly now.

BOINC is 99.9% fool proof if you leave it alone and don't mess with it.

If you get problems with it in the coming days, remove the whole BOINC manager program and download it from scratch.

EDIT; Keep an eye on your linux machine to see that it returns the next 2 or 3 work units ok.

John in Ireland.

---

Entertainingly, BOINC or SETI@home made a face as it died!

;)

---

It may not be 1984 but George Orwell sure did see the future . . .

I'd very much like this particular incident to
be a nightmare, gentlefolk. Perhaps I should
elaborate a bit more about what I was thinking
before I kill those tasks.

My first thought at the time was that these new images
in slot/0 MUST be bogus.

First, here's something to consider ...

% locate setiathome-
/home/boinc/BOINC/projects/setiathome.berkeley.edu/setiathome-5.12.i686-pc-linux-gnu
/home/boinc/BOINC/projects/setiathome.berkeley.edu/setiathome-5.12.i686-pc-linux-gnu.so
/home/boinc/BOINC/slots/0/setiathome-5.12.i686-pc-linux-gnu
/home/boinc/BOINC/slots/0/setiathome-5.12.i686-pc-linux-gnu.so
% (BTW, the above is the entire output.)

The projects/setiathome-* images have a Sept 22, 2006 date.
The slot/0 ones were created April 20, 2007. i.e. brand new.

And, yes, I did run "locate", "ls -la" and "ls -lac" on the located
directories before I deleted the files.

I wish I had made a tarball of the files or even compared them to one
another before I deleted them. So, much for forensic software investigation. What can I say? It was 2:45 AM.

What I WAS wondering was this: How did some clever S*B get the SETI@home
project to download these bogus images for him/her?!

And, no, it wasn't an alien. It was a smiley sticking out it's tongue. :-^

moved to number crunching.
you should get better response here.

---

With each crime and every kindness we birth our future.

I'm not a programmer, but my understanding is:

Boinc creates the slots (0, 1, 2, etc) to house temp data used by each thread for each project application result (task running/waiting). Got one processor and one project attached? you should have "slot 0" with stuff. If a wu has been suspended or "waiting to run" while another is runnning, then it goes in another slot (slot 1, for examplo). There can be empty slots, but the ones with stuff correspond to work it's doing or has started and will be back to doing.

I'm sure someone around here can explain it better.

I'm not a programmer, but my understanding is:

Boinc creates the slots (0, 1, 2, etc) to house temp data used by each thread for each project application result (task running/waiting). Got one processor and one project attached? you should have "slot 0" with stuff. If a wu has been suspended or "waiting to run" while another is runnning, then it goes in another slot (slot 1, for examplo). There can be empty slots, but the ones with stuff correspond to work it's doing or has started and will be back to doing.

I'm sure someone around here can explain it better.

Agreed, and once created the slots are never deleted - though they may be empty, as Tony says. My 8-core has 33 slot folders at the moment (0-32 inclusive): anyone claim anything higher ;-) ?

I don't understand the programming behind it either, but observation (Windows only) tells me:

There's always a file with the same name as the main application executable, but it's tiny: 1KB or smaller. Must be a pointer to the real thing in the projects directory.

Every project I'm running keeps some data files in there, but they're usually very small (or, again, pointers to real data elsewhere). Einstein is unusual in keeping the best part of a megabyte in there: Astropulse has 256KB: CPDN has 50 files, but they only add up to about 10KB between them.

Whatever you do, don't mess with anything in there unless you're prepared to lose a WU.

The slots directory(s) are the working directories and are full of files that are, for the most part, newly created 'soft links' to the original files... With a few exceptions...

I don't know why Boinc crashed, but I do know that when Boinc does crash it leaves behind files that it obviously can't clean up because, well, it crashed...

The pic I posted was of the Boinc icon when Boinc was disconnected from the local host, and most likely, would also make an appearance during a kill command being executed on the Boinc app... The smaller one in the menu-bar resembles a smilie face sticking its tongue out, but to me, it looks more like a funny monkey... When enlarged in the dock, it is clearly nothing more than a bulls-eye...

I don't think the original poster has anything to worry about other than why Boinc crashed in the first place... ;)

---

It may not be 1984 but George Orwell sure did see the future . . .