Thanks Philip.
One thing you didn't mention: are you using javac? If so, do you have
to trick it into working?
Particularly interested in hosts where you run Sun's JDK, since that's
the one we're using for now.
What happens when javac calls processes which return error codes? Does
javac report as successful, or is it accurate with its return codes?
Thanks again,
-m
Philip Senechal wrote:
>
> Not sure what help I can be with this, but we're running Java on RH7 with
> the following config:
>
> Glibc-2.2-12
> Apache 1.3.14-3
> ApacheJServ 1.1.2-1
> Sun JSDK 2.0
> IBM Java2 1.3 JDK/JVM
> or
> Sun Java2 1.3 JDK/JVM
>
> We've run under both IBM and Sun's JDK/JVM combos with no problems. We use
> IBM for single processor installs, and Sun for SMP since IBM doesn't support
> SMP. We're also using Sun's JSDK 2.0 because the one that comes with
> ApacheJServ is a bit buggy.
>
> The major difference we can see, is that IBM is strictly native
> threads...but the new Sun JDK/JVM is a mixed threads implementation.
>
> No real problems getting this installed...it all seemed to be snap.
>
> We're currently testing with RH Wolverine - 2.4.1 kernel and glibc-2.2.2-3.
> We did have to rollback to the i386 version of the glibc in order to get
> Java to work properly on Wolverine. The i686 (installed by default) version
> would not run. I'll let you know what we find.
>
> Hope that provided a little insight.
>
> Philip
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]]On Behalf Of Michael R. Jinks
> Sent: Wednesday, March 14, 2001 4:00 PM
> To: [EMAIL PROTECTED]
> Subject: java, RH7, and native threads
>
> We've started to roll out RH7 on newly-deployed machines. Recently
> we've discovered that the version of glibc shipped with RH7 (glibc-2.2)
> has a bug which causes javac to return 0 values even if subprocesses
> have encountered errors. A bug report in the JDK knowledge base lists
> "upgrade to Red Hat 7.1" as the solution.
>
> Well okay, so until 7.1 actually comes out, I thought I'd apply the
> latest glibc (2.2.2). That fixed the problem with javac, but broke java
> itself. From scratching my head over strace output, it seems as though
> java now uses native threads instead of green threads, that doesn't
> work, and java ends up stuck running nanosleep(2) forever, waiting for
> one of its threads to check in.
>
> So we've rolled back to glibc-2.2 and instructed all of our developers
> to run javac with the -classic argument, which seems to avoid the
> problems we're having but results in a significant performance hit.
>
> [pausing to breathe]
>
> Now my question: is anybody successfully using Java with native threads
> on RH7? Any hints you could pass along?
>
> --
> Michael Jinks, IB // Technical Entity // Saecos Corporation
> Opinions expressed above are my own, and not those of my employer.
>
> _______________________________________________
> Redhat-list mailing list
> [EMAIL PROTECTED]
> https://listman.redhat.com/mailman/listinfo/redhat-list
>
> _______________________________________________
> Redhat-list mailing list
> [EMAIL PROTECTED]
> https://listman.redhat.com/mailman/listinfo/redhat-list
--
Michael Jinks, IB // Technical Entity // Saecos Corporation
Opinions expressed above are my own, and not those of my employer.
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
