[DGD]Problem with state-dump

[E. Harte]

On Tue, 4 May 1999 balduin at uni-paderborn.de wrote:

> Hello !
> 
[...]
> - When I first did a dump_state() followed by a shutdown(),
>   I noticed after restart, that the connection objects
>   (cloned by connection() in e.g. the telnet port object to 
>    represent the connection) (oops many "connection"s in this sentence)
>   where still existent. I now close down all connections
>   when preparing dump_state() to avoid dumping them. But this
>   leaves me with the following error during shutdown()
> 
>   * sniplet from the logfile
>   [
>   shutting down ...
>   -----------------------------------------------------------------------
>   Tue May  4 12:14:06 1999, runtime error, ticks left: 3989945
>   Bad argument 1 for kfun call_other
>        103 receive_message    /kernel/net/user (#22)
>        198 handle_command     /system/user (#15)
>        150 command_loop       /system/user (#15)
>         66 input_to           /kernel/lib/auto (/system/user#15)
>   ]
> 
>   where this is the stuff in driver.c
> 
>   do_shutdown()
>   {
>       object *_u;
>       int i,sz;
>     
>       send_message("cutting connections ...\n");
>       _u = users();
>     
>       for (i=0,sz=sizeof(_u);i<sz;i++)
>       {
> 	  send_message("removing "+object_name(_u[i]));
>   	_u[i]->remove_object();
>       }
>       send_message("preparing dump ...\n");
>       prepare_dump();
>       send_message("dumping ...\n");
>       dump_state();
>       send_message("shutting down ...\n");
>       shutdown();
>   }

Does the 'shutdown' happen from a user command or is it done from a
separate thread, meaning from a 'call_out()'?  If the first option is
true, then it would be a good guess that our /system/user.c code does not
anticipate going linkdead halfway during a command-thread, if the second
option is true then I don't know why this is happening either. :-)

[...]
> - This leads to the next problem. I'd like to do a state-dump once in
>   a while, just to reduce my risk in case of emergency. Lets say shortage
>   of electricity. Or a server crash during development. Is there a way
>   to do this without saving/loosing the connection objects to avoid the 
>   object leak mentioned above ?

You could store the 'current' users() array in a global variable in the
driver object that is then saved in the 'periodical' statedump.  Make sure
the array is empty or nil if you do a shutdown + statedump as done in the
do_shutdown() code.  Then in restored(), check for the global variable and
make all user objects go linkdead after all.

A possible side-effect (which you can work around of course if you want)
is that the user-objects will seem to have been logged on from the time
the periodical statedump was made until the time that you restart from it,
which can look a bit silly (players aging a day while in reality they
logged out 5 minutes after the periodical statedump, things like that).

Hope that helps.

Erwin.


List config page:  http://list.imaginary.com/mailman/listinfo/dgd