RFR: 5049299 - (process) Use,posix_spawn, not fork, on S10 to avoid swap,exhaustion (jdk7u-dev)

[Volker Simonis]

Hi Amit,

I think this is the wrong place to ask. This list is about OpenJDK/OracleJDK.

As you can read from the bug report (better look at:
https://bugs.openjdk.java.net/browse/JDK-5049299) there's a list in
which OpenJDK/OracleJDK versions this bug has been fixed.
It describes that this bug hasn't been fixed in JDK 6 (only 7 and 8).

No idea where you can get this information for JRockit - you'd
probably have to ask the one from which you bought it.

By the way, in Oracle/OpenJDK 7 and 8 you can control the launch behaviour with:

-Djdk.lang.Process.launchMechanism={fork, vfork, posix_spawn}

Notice that not every platform supports every setting and different
platforms have different default launching mechanisms. You'll have to
look at the sources to see what's the default (i.e. the files
jdk/src/solaris/classes/java/lang/UNIXProcess.java.{bsd, linux,
solaris}').

Regards,
Volker



On Mon, Dec 29, 2014 at 3:51 PM, Amit Baxi <amit.baxi at agnity.com> wrote:
> Hi Folks,
>
> We are facing issue with process builder on JRockit JVM with solaris 5.10
> machine. That is realted to following bug:
>
>            http://bugs.java.com/view_bug.do?bug_id=5049299
>
>
>
> Currently we are using JRockit JDK 1.6.0_37-R28.2.5-4.1.0 version. Please
> let me know  following :
>
>  * Which version of JRockit this bug is fixed ? (as it is fixed in JAVA
>    8u 15)
>  * If yes then  is it also fixed for solaris 5.10 or In solaris process
>    builder still uses fork() call ?
>
>
> As per this page link below Mr. Rob wrote that for solaris it still using
> fork which is confusing weather fixed for solaris or not:
>
> http://mail.openjdk.java.net/pipermail/core-libs-dev/2013-November/022909.html
>
> Please confirm for this fix and thanks in advance.
>
> Waiting for your response,
>
> Thanks!
> *Amit Baxi
>
>
>
> *
>
>
>
>
>