Openjdk-infrastructure: test-godaddy-debian8-x64-3.adoptopenjdk.net not properly configured?

[jerboaa]

I'm seeing some test failures which look like test system setup issues:

test-godaddy-debian8-x64-3.adoptopenjdk.net: test-godaddy-debian8-x64-3.adoptopenjdk.net: Name or service not known

See:
https://ci.adoptopenjdk.net/view/Test_upstream/job/Test_openjdk11_hs_sanity.openjdk_x86-64_linux_upstream/36/testReport/junit/java_net_DatagramPacket_Constructor/java/Constructor/
https://ci.adoptopenjdk.net/view/Test_upstream/job/Test_openjdk11_hs_sanity.openjdk_x86-64_linux_upstream/36

Looks like a similar/same issue to #1134.

[sxa]

GoDaddy Debian machines disabled by @smlambert

[adam-thorpe]

@sxa555 Can we add test-godaddy-ubuntu1604-x64-1 and 3 to this list and disable those too? Reporting similar failures. 2 and 4 are already offline

[lumpfish]

Some system tests fail due to this configuration error too. Messages like:

10:14:24  LT  Test failed:
10:14:24  LT  java.net.UnknownHostException: test-godaddy-ubuntu1604-x64-3.adoptopenjdk.net: test-godaddy-ubuntu1604-x64-3.adoptopenjdk.net: Name or service not known
10:14:24  LT    at java.base/java.net.InetAddress.getLocalHost(InetAddress.java:1632)
10:14:24  LT    at gnu.testlet.java.net.DatagramPacket.DatagramPacketTest2.<init>(DatagramPacketTest2.java:70)
10:14:24  LT    at java.base/java.lang.J9VMInternals.newInstanceImpl(Native Method)
10:14:24  LT    at java.base/java.lang.Class.newInstance(Class.java:2089)
10:14:24  LT    at net.adoptopenjdk.loadTest.adaptors.MauveAdaptor.executeTest(MauveAdaptor.java:52)
10:14:24  LT    at net.adoptopenjdk.loadTest.LoadTestRunner$2.run(LoadTestRunner.java:182)
10:14:24  LT    at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1130)
10:14:24  LT    at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:630)
10:14:24  LT    at java.base/java.lang.Thread.run(Thread.java:835)
10:14:24  LT  Caused by: java.net.UnknownHostException: test-godaddy-ubuntu1604-x64-3.adoptopenjdk.net: Name or service not known
10:14:24  LT    at java.base/java.net.Inet6AddressImpl.lookupAllHostAddr(Native Method)
10:14:24  LT    at java.base/java.net.InetAddress$PlatformNameService.lookupAllHostAddr(InetAddress.java:932)
10:14:24  LT    at java.base/java.net.InetAddress.getAddressesFromNameService(InetAddress.java:1505)
10:14:24  LT    at java.base/java.net.InetAddress$NameServiceAddresses.get(InetAddress.java:851)
10:14:24  LT    at java.base/java.net.InetAddress.getAllByName0(InetAddress.java:1495)
10:14:24  LT    at java.base/java.net.InetAddress.getLocalHost(InetAddress.java:1627)
10:14:24  LT    ... 8 more
10:14:24  LT  <<<

[sxa]

@sxa555 Can we add test-godaddy-ubuntu1604-x64-1 and 3 to this list and disable those too? Reporting similar failures. 2 and 4 are already offline

Done

[sxa]

Side note to the above - test-godaddy-ubuntu1604-1 in particular seems to have performance issues (potentailly caused by timeouts of course) based on https://github.com/AdoptOpenJDK/openjdk-infrastructure/issues/1165

[M-Davies]

This is still affecting test-godaddy-debian8-x64-4.adoptopenjdk.net. See https://github.com/AdoptOpenJDK/openjdk-tests/issues/1750#issuecomment-619974621

[jerboaa]

@sxa We are seeing issues for test-godaddy- ubuntu1604-x64-4.adoptopenjdk.net. Same as https://github.com/AdoptOpenJDK/openjdk-tests/issues/1750#issuecomment-619974621 Failures are:

java.net.UnknownHostException: test-godaddy-ubuntu1604-x64-4.adoptopenjdk.net: test-godaddy-ubuntu1604-x64-4.adoptopenjdk.net: Name or service not known
    at java.net.InetAddress.getLocalHost(InetAddress.java:1506)
    at BadHandshakeTest.main(BadHandshakeTest.java:113)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:498)
    at com.sun.javatest.regtest.agent.MainActionHelper$AgentVMRunnable.run(MainActionHelper.java:298)
    at java.lang.Thread.run(Thread.java:748)
Caused by: java.net.UnknownHostException: test-godaddy-ubuntu1604-x64-4.adoptopenjdk.net: Name or service not known
    at java.net.Inet6AddressImpl.lookupAllHostAddr(Native Method)
    at java.net.InetAddress$2.lookupAllHostAddr(InetAddress.java:929)
    at java.net.InetAddress.getAddressesFromNameService(InetAddress.java:1324)
    at java.net.InetAddress.getLocalHost(InetAddress.java:1501)
    ... 7 more

[jerboaa]

https://ci.adoptopenjdk.net/job/Test_openjdk8_hs_sanity.openjdk_x86-64_linux_upstream/42/

[M-Davies]

@jerboaa Those test failures are being excluded in https://github.com/AdoptOpenJDK/openjdk-tests/pull/1785 after triaging in https://github.com/AdoptOpenJDK/openjdk-tests/issues/1750

[jerboaa]

@M-Davies Sure, but they work fine on a properly configured system. Shame they need to be excluded because of this.

[smlambert]

Either the tests need to be temporarily excluded until the machine config can be corrected or this machine needs to be removed from the test role (remove ci.role.test label) until the machine config can be corrected, so that we do not keep hitting it and failing the tests. I am fine with either approach.

[sxa]

Marked godaddy-ubuntu-4 offline as per Severin's comment ... If people wouldn't dump machines live into my infrastructure and then ignore failures that would be the best way to keep things stable...

[Willsparker]

https://github.com/AdoptOpenJDK/openjdk-infrastructure/issues/1423#issuecomment-654241669

I managed to find a solution to an issue that has the same error message as listed here, therefore it's worth trying to apply the fix to the other GoDaddy machines and see if the tests that were excluded now run on them.
For the sake of these, I'm going to just use the com/sun/jdi/BadHandshakeTest.java test that was excluded with https://github.com/AdoptOpenJDK/openjdk-tests/pull/1785

| Machine | Grinder Run | Pass? |
|------------------|-------------|-------|
| ubuntu1604-x64-1 | 3531 | yes |
| ubuntu1604-x64-2 | 3518 | yes |
| ubuntu1604-x64-3 | 3519 | yes |
| ubuntu1604-x64-4 | 3520 | yes |
| debian8-x64-1 | 3528 | yes |
| debian8-x64-2 | 3527 | yes |
| debian8-x64-3 | 3529 | yes |
| debian8-x64-4 | 3530 | yes |

• debian8-x64-1 is currently disabled (but still connected to Jenkins) as @gdams marked it as 'flaky' in May. I enabled it to run the grinder, and disabled it again with the same message.

The Debian8 machines seem to already have the fix applied to the /etc/hosts file, but not the template file so they would fail after a restart, so I rectified that.

Just to confirm these weren't false positives too, I commented out the fix in /etc/hosts for Ubuntu1604-x64-1 and reran the grinder and it failed with the hostname error message.

It seems the issue is fixed :-) @M-Davies Can we un-exclude those tests now?

[M-Davies]

PR is in to do that here: https://github.com/AdoptOpenJDK/openjdk-tests/pull/1870

[Willsparker]

With the above PR merged, I think this can be closed now.