MOO-cows Mailing List Archive
[Prev][Next][Index][Thread]

Re: DB crasher

> On a MOO that I'm on, a "clock" runs in the background.  This monitors the
> number of tasks active on the MOO.  If it exceeds a certain number, all
> currently active tasks (except the clock) are killed and the owners of the
> tasks are sent mail about the fact that their tasks just got toasted.
> Unfortunately, the MOO is starting to have more players, and the clock is
> erroneously assuming a fork bomb is in progress.  Worse, the guy who wrote it
> has left the wizard staff leaving wizards who don't want to touch it for fear
> that it will blow up in their faces.
> 
> -manta (NOT a wizard of that MOO)

Tell the wizzes of that moo to upgrade to 1.7.9p2 or the latest 1.8.0, 
because that already has a builtin task number monitor that's much more 
effecient than such a clock and won't have the problem of erroniously 
killing processes, because it simply won't let anyone begin any new 
processes over their task quota limit.  Although it's their moo so it's 
their call, running a clock like that is an _extremely_ messy band-aid 
fix to an old problem that doesn't exist anymore.  What's worse, the 
chance that the moo will crash and loose data since the last checkpoint, 
or the 100% certainty that all processes will suddenly crash and almost 
certainly loose data in one way or another?  I'd opt to simply let it crash.

(Note that, if there really were a true fork bomb attack on that moo, 
unless the clock cycles so frequently as to lag the entire moo, the fork 
bomb could grow so fast that the clock will choke and die trying to 
keep up..  If they don't believe me, tell them to give me a prog bit on 
their moo and an okay to go ahead and crash it.  Whoever wrote it 
must not understand exponential mathematics.)

Phantom
References:
Home | Subject Index | Thread Index