Backgroundrb devel - Nov 2006 - Zombies?

Hey all,

Quick question about handling completed workers...

Most of my workers are one-offs that just let me spin off a long-running
file transfer process and then they just need to self-destruct when
completed. Thus, at the end of my do_work, I just call self.delete to (in
theory) self-destruct.

However, while checking the jobs.size from the console, I''ve noticed
that
this doesn''t seem to be working. Further, by grabbing the worker keys
via
jobs.keys, I try to manually call delete_worker to kill them off, and I get
a RuntimeError "not beating" from Slave. However, when I call
jobs[key].shutdown? to see if the job is dead, it returns false, indicating
the the job isn''t "shut down". So, I can''t seem to
have the workers
self-destruct, and I can''t kill them from the MiddleMan, either. The
only
solution seems to be to call shutdown({:quiet => true}) to suppress errors
from slave. However, I don''t know if this will actually succeed in
whacking
the process...

Calling gc! basically results in the same behavior, because it tries to call
delete_worker and I get the same "not beating" error.

If this is a bug, I''ll put it in trac, but it seems like I must be
doing
something wrong.

Thanks!

-- 
Thermal Creative
http://blog.thermalcreative.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://rubyforge.org/pipermail/backgroundrb-devel/attachments/20061123/f613ecac/attachment.html

ok, this was was lingering for a while as I suspected it was related to
another issue, but if you are running off trunk, the fix for this is in
as of:

  http://backgroundrb.devjavu.com/projects/backgroundrb/changeset/160

we shouldn''t be that far off from 0.2.1 now.

/skaar


* Matt White (stockliasteroid at gmail.com) [061122
16:47]:
>    Hey all,
> 
>    Quick question about handling completed workers...
> 
>    Most of my workers are one-offs that just let me spin off a long-running
>    file transfer process and then they just need to self-destruct when
>    completed. Thus, at the end of my do_work, I just call self.delete to
(in
>    theory) self-destruct.
> 
>    However, while checking the jobs.size from the console, I''ve
noticed that
>    this doesn''t seem to be working. Further, by grabbing the
worker keys via
>    jobs.keys, I try to manually call delete_worker to kill them off, and I
>    get a RuntimeError "not beating" from Slave. However, when I
call
>    jobs[key].shutdown? to see if the job is dead, it returns false,
>    indicating the the job isn''t "shut down". So, I
can''t seem to have the
>    workers self-destruct, and I can''t kill them from the
MiddleMan, either.
>    The only solution seems to be to call shutdown({:quiet => true}) to
>    suppress errors from slave. However, I don''t know if this will
actually
>    succeed in whacking the process...
> 
>    Calling gc! basically results in the same behavior, because it tries to
>    call delete_worker and I get the same "not beating" error.
> 
>    If this is a bug, I''ll put it in trac, but it seems like I must
be doing
>    something wrong.
> 
>    Thanks!
> 
>    --
>    Thermal Creative
>    [1]http://blog.thermalcreative.com
> 
> References
> 
>    Visible links
>    1. http://blog.thermalcreative.com/
> _______________________________________________
> Backgroundrb-devel mailing list
> Backgroundrb-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/backgroundrb-devel
-- 
----------------------------------------------------------------------
|\|\             where in the       |          s_u_b_s_t_r_u_c_t_i_o_n 
| | >===========  W.A.S.T.E.        |                  genarratologies 
|/|/    (_)     is the wisdom       |                  skaar at waste.org
----------------------------------------------------------------------