win7 jdk 8 build : fork: retry: Resource temporarily unavailable

Phil Race philip.race at oracle.com
Thu Aug 22 20:30:52 UTC 2013


On 8/22/2013 1:12 AM, Erik Joelsson wrote:
> "fork: retry: Resource temporarily unavailable" is a classic cygwin 
> problem. I don't think it's related to amount of memory used. One 
> thing you can try is rebaseall described here: 
> http://cygwin.wikia.com/wiki/Rebaseall. It sometimes helps.
>


 >Sometimes, after updating or installing packages, you'll start to get 
strange errors related
 > to "fork()" or .dll loading. These errors are usually solved by 
rebasing your packages.

I remember doing this several years ago (and it was a bit fraught). Then 
it made sense
because I had done some updates and I had trouble running cygwin (the 
shell I think).
Here I've not done anything like that and I cured it by quitting a few apps.
I think I need to wait for the next time it occurs to try this to see if 
it really will help.

-phil.

> /Erik
>
> On 2013-08-21 20:56, Phil Race wrote:
>> My Win 7 x 64 showed it had approx 2GB (out of 6GB) memory free when 
>> I got this message
>> doing an incremental build :-
>>
>> ## Starting langtools
>> /cygdrive/c/jdks/jdk8_2d/common/bin/logger.sh: fork: retry: Resource 
>> temporarily unavailable
>> /cygdrive/c/jdks/jdk8_2d/common/bin/logger.sh: fork: retry: Resource 
>> temporarily unavailable
>> /cygdrive/c/jdks/jdk8_2d/common/bin/logger.sh: fork: retry: Resource 
>> temporarily unavailable
>> /cygdrive/c/jdks/jdk8_2d/common/bin/logger.sh: fork: retry: Resource 
>> temporarily unavailable
>> /cygdrive/c/jdks/jdk8_2d/common/bin/logger.sh: fork: Resource 
>> temporarily unavailable
>> make: *** [langtools-only] Error 254
>>
>> I wonder if something like this is responsible for full builds often 
>> needing several tries to
>> complete as well, even though the errors there seem somewhat random.
>> If you see below, the build has decided it might use up to 4Gb (if I 
>> read this right)
>> which seems  like a lot for a system with only(!) 6GB RAM. But surely 
>> it would not
>> hit this limit right at the outset ?
>>
>> Build config info follows. Its an 8 core Xeon processor ..
>>
>> Configuration summary:
>> * Debug level:    release
>> * JDK variant:    normal
>> * JVM variants:   server,client
>> * OpenJDK target: OS: windows, CPU architecture: x86, address length: 32
>>
>> Tools summary:
>> * Environment:    cygwin version 1.7.8(0.236/5/3) (root at 
>> /cygdrive/c/cygwin)
>> * Boot JDK:       java version "1.7.0_40-ea"  Java(TM) SE Runtime 
>> Environment (build 1.7.0_40-ea-b35)  Java HotSpot(TM) 64-Bit Server 
>> VM (build 24.0-b54, mixed mode)   (at 
>> /cygdrive/c/progra~1/java/jdk17~1.0_4)
>> * C Compiler:     Microsoft CL.EXE version 16.00.40219.01 (at 
>> /cygdrive/c/vs2010/VC/BIN/cl)
>> * C++ Compiler:   Microsoft CL.EXE version 16.00.40219.01 (at 
>> /cygdrive/c/vs2010/VC/BIN/cl)
>>
>> Build performance summary:
>> * Cores to use:   3
>> * Memory limit:   4095 MB
>> * ccache status:  not available for your system
>>
>> -phil.
>>
>>




More information about the build-dev mailing list