[erlang-questions] Weird pg2 behavior

[Matthew Dempsky]

For what it's worth, this is on a network of R12B-1 nodes running on
Linux amd64 and FreeBSD i386.

On Thu, Apr 3, 2008 at 7:24 PM, Matthew Dempsky <matthew@REDACTED> wrote:
> We're seeing weird behavior with pg2, namely that dead pids are not
>  getting cleared from the pg2:get_members table.  E.g.
>
>  250> [X || X <- pg2:get_members(Group) --
>  pg2:get_local_members(Group), node(X) =:= node()].
>  [<0.795.0>,<0.794.0>,<0.793.0>,<0.792.0>,<0.791.0>,
>   <0.790.0>,<0.789.0>,<0.773.0>]
>
>  As I understand pg2, the above expression should always evaluate to an
>  empty list (ignoring race conditions).  I've looked over pg2's source
>  code briefly, and I don't see anything obvious that could cause this.
>
>  Am I correct that this is a bug in pg2?  Can anyone suggest what might
>  cause this?
>