#5476: Cap memory usage for one thread without disrupting other threads
---------------------------------+------------------------------------------
Reporter: joeyadams | Owner:
Type: feature request | Status: new
Priority: normal | Component: Runtime System
Version: 7.2.1 | Keywords:
Testcase: | Blockedby:
Os: Unknown/Multiple | Blocking:
Architecture: Unknown/Multiple | Failure: None/Unknown
---------------------------------+------------------------------------------
I would like to be able to programmatically guard against runaway memory
usage by a thread. If it occurs, I want to be able to handle it
gracefully, and I do not want other threads to be disrupted.
Note: I [http://stackoverflow.com/questions/7297654/how-to-cap-memory-
usage-of-haskell-threads asked this on StackOverflow] a few days ago. The
paragraph below is copied from the question.
For example, suppose I want to write a server, hosting a scripting
language interpreter, that users can connect to. It's Turing-complete, so
programs could theoretically use unlimited memory or time. Suppose each
client is handled with a separate thread. If a client writes an infinite
loop that consumes memory very quickly, I want to ensure that the thread
consumes no more than, say, 1 MB of memory, before being alerted with an
exception. I do not want other users to be affected when that happens.
Ideally, it would be a function like this:
{{{
catchHeapOverflow :: Int64 -> IO a -> IO a -> IO a
catchHeapOverflow limitBytes handler action = ...
}}}
It sets a memory usage limit for the current thread (and any child
threads, ideally). If this limit is surpassed while performing the
action, it invokes the handler (with the usage limit lifted).
Granted, the notion of a thread using too much memory makes less sense
when objects are shared between threads. Should a thread be blamed for
pointing to a lot of memory that another thread gave it? I don't know.
Should a thread be blamed for consuming a lot of memory while evaluating a
thunk? Absolutely. In my opinion, handling the edge cases perfectly is
not as important as being able to do this at all. As long as the user
performs all the necessary calculations in one thread and does not export
unforced thunks to other threads, s/he should be fine.
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5476>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
