RFR(trivial): 8204529 gc/TestAllocateHeapAtMultiple.java fail with Agent 7 timed out

Erik Joelsson erik.joelsson at oracle.com
Fri Nov 9 01:29:07 UTC 2018


Change looks good to me.

It should be noted that there is some underlying issue that also needs 
to be investigated since the test normally runs in a few seconds and 
only for specific jvm args and types of machines (newer Solaris sparc) 
does it go upwards of 18m. A separate issue should be filed for that.

/Erik


On 2018-11-08 17:00, Daniel D. Daugherty wrote:
> Greetings,
>
> Erik J. has done a bunch of analysis on the timeout failures for
> this test: gc/TestAllocateHeapAtMultiple.java. He has determined
> that increasing the default timeout value will help this test
> pass on the machines where it was timing out.
>
> JDK-8204529 gc/TestAllocateHeapAtMultiple.java fail with Agent 7 timed 
> out
> https://bugs.openjdk.java.net/browse/JDK-8204529
>
> along with a bunch of analysis for that test in:
>
> JDK-8205526 tests timeout intermittently in "agentvm" mode
> https://bugs.openjdk.java.net/browse/JDK-8205526
>
>
> Here's the context diff for the trivial fix:
>
> $ hg diff
> diff -r 3e0ebf913679 
> test/hotspot/jtreg/gc/TestAllocateHeapAtMultiple.java
> --- a/test/hotspot/jtreg/gc/TestAllocateHeapAtMultiple.java    Thu Nov 
> 08 19:16:16 2018 -0500
> +++ b/test/hotspot/jtreg/gc/TestAllocateHeapAtMultiple.java    Thu Nov 
> 08 19:51:21 2018 -0500
> @@ -1,5 +1,5 @@
>  /*
> - * Copyright (c) 2017, Oracle and/or its affiliates. All rights 
> reserved.
> + * Copyright (c) 2017, 2018, Oracle and/or its affiliates. All rights 
> reserved.
>   * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
>   *
>   * This code is free software; you can redistribute it and/or modify it
> @@ -27,6 +27,7 @@
>   * @library /test/lib
>   * @modules java.base/jdk.internal.misc
>   * @requires vm.bits == "64" & vm.gc != "Z" & os.family != "aix"
> + * @run main/timeout=360 TestAllocateHeapAtMultiple
>   */
>
>  import jdk.test.lib.JDKToolFinder;
>
>
> Thanks, in advance, for any comments, suggestions or questions.
>
> Dan
>



More information about the hotspot-runtime-dev mailing list