Intermittent ZipExceptions in Linux/s390x openjdk tests

[sxa]

many test targets failing in sanity and extended openjdk with ZipException

This is interesting ... It's not consistent. Here are the results from the jdk_security1_1 target on various systems/versions:

Version	Machine	Result
11.0.20.1	test-docker-sles15-1	2 failures
11.0.20.1	test-marist-sles15-2	Pass
11.0.20.1	test-marist-ubuntu2204-1	1 failure
11.0.20	test-docker-sles15-1	1 failure
17.0.8.1	test-docker-sles12-s390x-1	Pass
17.0.8.1	test-docker-sles15-1	Pass
11.0.20.1	test-marist-sles15-2	Pass (10 iterations)
17.0.8.1	test-marist-ubuntu2204-1	Pass (20 iterations)
17.0.8.1	test-docker-sles15-1	Pass (20 iterations)
11.0.20.1	test-marist-ubuntu2204-1	Failed 9/10
11.0.20.1	test-docker-sles15-1	Failed 9/10

On the basis of the above I'm re-running the failed suites on test-marist-sles15-2:

• sanity failures: Grinder 7500: jdk_lang_0,jdk_lang_1,jdk_security1_0,jdk_security1_1,jdk_security2_0,jdk_security2_1,jdk_util_0,jdk_util_1,jdk11_tier1_pack200_0,jdk11_tier1_pack200_1 (Passed up to jdk_util a the time of looking, so it's positive)
• extended failures: Grinder 7501: jdk_security3_1,jdk_beans_1,jdk_security3_0,jdk_tools_0,jdk_imageio_1,jdk_tools_1,jdk_imageio_0

[sxa]

We need to understand what's different here, particularly between the SLES15 systems https://ci.adoptium.net/computer/test-docker-sles15-s390x-1 and https://ci.adoptium.net/computer/test-marist-sles15-s390x-2 which have reasonable repeatable differences in behavior despite both being SLES15. The docker one is hosted on https://ci.adoptium.net/manage/computer/dockerhost-marist-ubuntu2204-s390x-1 so it's not inconceivable that this is related to a kernel issue of some sort which would be different between the two.

[sxa]

The extended Grinder 7501 is not going well - lots of timeouts occurring such as

14:51:58  ACTION: main -- Error. Program `/home/jenkins/workspace/Grinder_testList_0/openjdkbinary/j2sdk-image/bin/java' timed out (timeout set to 960000ms, elapsed time including timeout handling was 960549ms).
14:51:58  REASON: User specified action: run main/othervm RespondToRetransmit client 0 hello_request 
14:51:58  TIME:   960.55 seconds
14:51:58  messages:

Attempting to run on other systems:

• G#7507: ZipExceptions on test-docker-sles15-s390x-1
• G#7508: ZipExceptions on test-marist-rhel8-s390x-2
• G#7509: ZipExceptions on test-docker-sles12-s390x-1

[sxa]

RHEL8 run is not the same reflection errors that the other two are seeing (and jdk_beans passed on that one)
Temurin Compliance Grinder 3714 (RHEL7) and 3715 (ubuntu22.04-1) are failing with ZipException messages and timeouts. 3716 on the ubuntu2004-3 machine looks to be passing (Not yet finished) so trying on an Ubuntu 20.04 on the main server at https://ci.adoptium.net/job/Grinder/7511/

EDIT: Ubuntu20.04 passed all of the extended subset listed in the description other than tools/launcher/JliLaunchTest.java.JliLaunchTest which failed with Use -nativepath to specify the location of native code in the Stack Trace output