Re: [GHC] #1840: Reduce strain on PPC Mac OS X swapspace when test case outofmem2 fails

[GHC]

#1840: Reduce strain on PPC Mac OS X swapspace when test case outofmem2 fails
-+--
 Reporter:  thorkilnaur  |  Owner: 
 Type:  bug  | Status:  closed 
 Priority:  normal   |  Milestone:  6.8.2  
Component:  Test Suite   |Version:  6.8.1  
 Severity:  normal   | Resolution:  fixed  
 Keywords:   | Difficulty:  Unknown
 Testcase:  outofmem2|   Architecture:  Unknown
   Os:  Unknown  |  
-+--
Changes (by simonmar):

  * status:  new => closed
  * resolution:  => fixed

Comment:

 I believe I've fixed this.  outofmem and outofmem2 now set a ulimit before
 running.  Also, the 6.8 branch had an incomplete fix for outofmem2, which
 I'm testing now.

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #1840: Reduce strain on PPC Mac OS X swapspace when test case outofmem2 fails

[GHC]

#1840: Reduce strain on PPC Mac OS X swapspace when test case outofmem2 fails
-+--
 Reporter:  thorkilnaur  |  Owner: 
 Type:  bug  | Status:  new
 Priority:  normal   |  Milestone:  6.8.2  
Component:  Test Suite   |Version:  6.8.1  
 Severity:  normal   | Resolution: 
 Keywords:   | Difficulty:  Unknown
 Testcase:  outofmem2|   Architecture:  Unknown
   Os:  Unknown  |  
-+--
Changes (by igloo):

  * milestone:  => 6.8.2

Comment:

 Thanks for the patch! We'll look into it, and the ulimit/overcommit
 problem, ASAP.

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

Re: [GHC] #1840: Reduce strain on PPC Mac OS X swapspace when test case outofmem2 fails

[GHC]

#1840: Reduce strain on PPC Mac OS X swapspace when test case outofmem2 fails
-+--
 Reporter:  thorkilnaur  |  Owner: 
 Type:  bug  | Status:  new
 Priority:  normal   |  Milestone: 
Component:  Test Suite   |Version:  6.8.1  
 Severity:  normal   | Resolution: 
 Keywords:   | Difficulty:  Unknown
 Testcase:  outofmem2|   Architecture:  Unknown
   Os:  Unknown  |  
-+--
Comment (by duncan):

 Also, on 64bit platforms with VM overcommit (like linux on amd64) this
 test fails by timing out. It allocates gigabytes and gigabytes of virtual
 address space but the OS never refuses the requests for more (until
 perhaps the whole of physical memory fills up with page tables). The test
 should set a ulimit to limit the amount of virtual address space the
 process is allowed. Then it'd actually work on standard 64bit linux and if
 the limit were set low enough then it'd work nicely on MacOSX which
 apparently likes to make sure there's as much swap space as allocated
 virtual address space.

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs

[GHC] #1840: Reduce strain on PPC Mac OS X swapspace when test case outofmem2 fails

[GHC]

#1840: Reduce strain on PPC Mac OS X swapspace when test case outofmem2 fails
+---
Reporter:  thorkilnaur  |   Owner:   
Type:  bug  |  Status:  new  
Priority:  normal   |   Milestone:   
   Component:  Test Suite   | Version:  6.8.1
Severity:  normal   |Keywords:   
  Difficulty:  Unknown  |Testcase:  outofmem2
Architecture:  Unknown  |  Os:  Unknown  
+---
 The outofmem2 test verifies that the runtime system reacts suitably to a
 request to allocate an amount of memory that exceeds the available heap
 space. See trac ticket #1791 for additional details.

 As a practical matter, when outofmem2 fails on my PPC Mac, it does so by
 succeeding to allocate about 1 Gbyte of memory. This causes a similar
 amount of additional disk space worth of swapfiles to be created. And this
 strains the capacity of my somewhat modest equipment.

 The attached patch supplies a suggested remedy for this, by simply
 reducing the amount ofmemory that outofmem2.hs attempts to allocate. It
 would be very useful for me to have this patch (or any other cure of the
 problem) also merged to the current stable where the problem occurs
 routinely.

 Best regards
 Thorkil

-- 
Ticket URL: 
GHC 
The Glasgow Haskell Compiler___
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs