Many network tests failing on AIX in JDK20 (with UnknownHostException: Hostname and service name not provided) #3178

smlambert · 2023-07-17T13:12:25Z

sun/security/krb5/auto/NoAddresses.java appears to fail across all available AIX machines:

java.net.UnknownHostException: adopt06: adopt06: Hostname and service name not provided or found
	at java.base/java.net.InetAddress.getLocalHost(InetAddress.java:1791)
	at NoAddresses.main(NoAddresses.java:51)
	at java.base/jdk.internal.reflect.DirectMethodHandleAccessor.invoke(DirectMethodHandleAccessor.java:104)
	at java.base/java.lang.reflect.Method.invoke(Method.java:578)
	at com.sun.javatest.regtest.agent.MainWrapper$MainThread.run(MainWrapper.java:125)
	at java.base/java.lang.Thread.run(Thread.java:1623)
Caused by: java.net.UnknownHostException: adopt06: Hostname and service name not provided or found
	at java.base/java.net.Inet6AddressImpl.lookupAllHostAddr(Native Method)
	at java.base/java.net.Inet6AddressImpl.lookupAllHostAddr(Inet6AddressImpl.java:52)
	at java.base/java.net.InetAddress$PlatformResolver.lookupByName(InetAddress.java:1061)
	at java.base/java.net.InetAddress.getAddressesFromNameService(InetAddress.java:1683)
	at java.base/java.net.InetAddress$NameServiceAddresses.get(InetAddress.java:1004)
	at java.base/java.net.InetAddress.getAllByName0(InetAddress.java:1673)
	at java.base/java.net.InetAddress.getLocalHost(InetAddress.java:1786)
	... 5 more

JavaTest Message: Test threw exception: java.net.UnknownHostException: adopt06: adopt06: Hostname and service name not provided or found
JavaTest Message: shutting down test

STATUS:Failed.`main' threw exception: java.net.UnknownHostException: adopt06: adopt06: Hostname and service name not provided or found

Test Info
Test Name: jdk_security4_1
Test Duration: 13 min 8 sec
Machine: test-osuosl-aix72-ppc64-6
TRSS link for the test output: https://trss.adoptium.net/output/test?id=64b3f9a817052c671580a5fe

Build Info
Build Name: Test_openjdk20_hs_sanity.openjdk_ppc64_aix
Jenkins Build start time: Jul 15 2023, 09:08 pm
Jenkins Build URL: https://ci.adoptium.net/job/Test_openjdk20_hs_sanity.openjdk_ppc64_aix/110/
TRSS link for the build: https://trss.adoptium.net/allTestsInfo?buildId=64b3f8a617052c671580a04d

Java Version
openjdk version "20.0.1-beta" 2023-04-18
OpenJDK Runtime Environment Temurin-20.0.1+9-202307152344 (build 20.0.1-beta+9-202307152344)
OpenJDK 64-Bit Server VM Temurin-20.0.1+9-202307152344 (build 20.0.1-beta+9-202307152344, mixed mode)

This test has been failed 17 times since Jun 10 2023, 08:34 pm
Java Version when the issue first seen
openjdk version "20.0.1" 2023-04-18
OpenJDK Runtime Environment Temurin-20.0.1+9 (build 20.0.1+9)
OpenJDK 64-Bit Server VM Temurin-20.0.1+9 (build 20.0.1+9, mixed mode)
Jenkins Build URL: https://ci.adoptium.net/job/Test_openjdk20_hs_sanity.openjdk_ppc64_aix/94/

The test failed on machine test-osuosl-aix72-ppc64-6 2 times
The test failed on machine test-osuosl-aix72-ppc64-3 3 times
The test failed on machine test-osuosl-aix72-ppc64-5 4 times
The test failed on machine test-osuosl-aix72-ppc64-4 2 times
The test failed on machine test-osuosl-aix72-ppc64-1 1 times
The test failed on machine test-osuosl-aix72-ppc64-2 4 times
The test failed on machine build-osuosl-aix72-ppc64-2 1 times

Rerun in Grinder

Other network related targets also failing on AIX with same issue, examples:
jdk_nio from https://ci.adoptium.net/job/Test_openjdk20_hs_extended.openjdk_ppc64_aix_testList_2/6/

The text was updated successfully, but these errors were encountered:

smlambert · 2023-09-05T11:48:14Z

Transferring this issue to infrastructure repository.

It continues to be an issue seen on certain machines, including test-osuosl-aix72-ppc64-5, as seen in https://ci.adoptium.net/job/Test_openjdk21_hs_extended.openjdk_ppc64_aix_testList_2/7/

04:05:27  STDERR:
04:05:27  java.net.UnknownHostException: adopt05: adopt05: Hostname and service name not provided or found
04:05:27  	at java.base/java.net.InetAddress.getLocalHost(InetAddress.java:1936)
04:05:27  	at jdk.test.lib.Utils.getHostname(Utils.java:450)
04:05:27  	at JstatdTest.getDestination(JstatdTest.java:111)
04:05:27  	at JstatdTest.runJps(JstatdTest.java:132)
04:05:27  	at JstatdTest.runToolsAndVerify(JstatdTest.java:209)
04:05:27  	at JstatdTest.runTest(JstatdTest.java:346)
04:05:27  	at JstatdTest.doTest(JstatdTest.java:314)
04:05:27  	at TestJstatdPortAndServer.main(TestJstatdPortAndServer.java:40)
04:05:27  	at java.base/jdk.internal.reflect.DirectMethodHandleAccessor.invoke(DirectMethodHandleAccessor.java:103)
04:05:27  	at java.base/java.lang.reflect.Method.invoke(Method.java:580)
04:05:27  	at com.sun.javatest.regtest.agent.MainActionHelper$AgentVMRunnable.run(MainActionHelper.java:333)
04:05:27  	at java.base/java.lang.Thread.run(Thread.java:1583)

smlambert · 2023-09-05T11:49:53Z

Related: #3030

aixtools · 2023-09-05T13:38:44Z

All hosts - or some? As most systems are a clone of adopt10.

I can look for differences - but a known working server compared with a known failing server works is the preferred starting point.

smlambert · 2023-09-05T17:44:52Z

I believe the issue is observed on test-osuosl-aix72-ppc64-4, test-osuosl-aix72-ppc64-5, test-osuosl-aix72-ppc64-6.

Though I have exhaustively looked at all hosts, anecdotally -1, -2, -3 look like they do not have this issue.

smlambert · 2023-09-05T17:49:43Z

To reproduce this issue, use this Rerun in Grinder link and set LABEL to be the hostname to run on, for example set LABEL to test-osuosl-aix72-ppc64-6.

adamfarley · 2023-09-06T09:36:42Z

It looks like -3 and -4 were removed from the ansible inventory file back in May, so perhaps these machines can be ignored if they are pending deletion.

I mention the pending deletion of these machines here.

aixtools · 2023-09-06T10:21:45Z

Here are the current records:

BUILD:
      - osuosl:
          aix72-ppc64-1: {ip: 140.211.9.166, description: p8-java1-adopt10.osuosl.org, 7200-02-04-1914}
          aix72-ppc64-2: {ip: 140.211.9.12, description: p8-aix1-adopt02.osuosl.org, 7200-02-04-1914}

TEST:
      - osuosl:
          aix72-ppc64-1: {ip: 140.211.9.28, description: p8-aix1-adopt03.osuosl.org, 7200-04-02-2028}
          aix72-ppc64-2: {ip: 140.211.9.36, description: p8-aix1-adopt04.osuosl.org, 7200-02-05-1938}
          aix72-ppc64-3: {ip: 140.211.9.168, description: p8-java1-adopt07.osuosl.org, 7200-02-04-1914}
          aix72-ppc64-4: {ip: 140.211.9.169, description: p8-java1-adopt08.osuosl.org, 7200-02-04-1914}
          aix72-ppc64-5: {ip: 140.211.9.99, description: p9-aix1-adopt05.osuosl.org, 7200-02-04-1914}
          aix72-ppc64-6: {ip: 140.211.9.100, description: p9-aix1-adopt06.osuosl.org, 7200-02-04-1914}
          aix73-ppc64-1: {ip: 140.211.9.10, description: p8-aix1-adopt01.osuosl.org, 7300-01-02-2320}

All systems are build and active.
-1 and -2 were built long ago (before the new build-aix72-1), the systems -[3,4,5,6} are all cloned from the build-aix72-1 system, the aix73 was built from DVD.

After the cloning the playbook, afaik, was rerun over all the all the aix72 systems - and obviously, the aix73 system was built/configured using the playbook.

Any differences are because someone has made changes manually.

No control and/or change history on manual changes.

aixtools · 2023-09-06T10:22:36Z

It looks like -3 and -4 were removed from the ansible inventory file back in May, so perhaps these machines can be ignored if they are pending deletion.

I mention the pending deletion of these machines here.

There are no aix71 system remaining - those are the systems that were removed in May.

adamfarley · 2023-09-06T11:31:56Z

There are no aix71 system remaining - those are the systems that were removed in May.

Oh, ok. Odd that the removed machines had "aix72" in their names.

adamfarley · 2023-09-06T11:46:06Z

Ok, got my facts straight now.

test-osuosl-aix72-ppc64-3 and -4 were removed from the inventory file in may, but they were later replaced by other machines that now use the same names as the ones that were removed.

So the current test-osuosl-aix72-ppc64-3 and -4 are not pending deletion from jenkins.

aixtools · 2023-09-06T11:51:37Z

Correct on both points. ***@***.***> Michael Felt Mobile +31 (0)6 5184 4181 Email ***@***.*** From: Adam Farley ***@***.***> Sent: Wednesday, September 6, 2023 1:46 PM To: adoptium/infrastructure ***@***.***> Cc: Michael Felt ***@***.***>; Comment ***@***.***> Subject: Re: [adoptium/infrastructure] Many network tests failing on AIX in JDK20 (with UnknownHostException: Hostname and service name not provided) (Issue #3178) Ok, got my facts straight now. test-osuosl-aix72-ppc64-3 and -4 were removed from the inventory file in may <#3053> , but they were later <e9c5559> replaced by other machines that now use names as the ones that were removed. So the current test-osuosl-aix72-ppc64-3 and -4 are not pending deletion from jenkins. — Reply to this email directly, view it on GitHub <#3178 (comment)> , or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACSZR5M7IWNFSINGGKPFOKTXZBPATANCNFSM6AAAAAA4LU5B4E> . You are receiving this because you commented. <https://github.com/notifications/beacon/ACSZR5O26T3XANYNTZUUL4DXZBPATA5CNFSM6AAAAAA4LU5B4GWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTTF2DORY.gif> Message ID: ***@***.*** ***@***.***> >

adamfarley · 2023-10-13T15:42:41Z

Update: This issue still appears to occur. Example.

java.net.UnknownHostException: adopt07: adopt07: Hostname and service name not provided or found

Seen on test-osuosl-aix72-ppc64-3.

sxa · 2023-11-02T11:08:53Z

NOTE:

I'm running Grinders 7992 through 7999 with some known problematic suites on different AIX 7.2 machines - can look at a summary of how it went with curl -s https://ci.adoptium.net/job/Grinder/7992/consoleText | egrep 'Running test |TEST RESULT: ' against each one. Some of them are showing the problems described above. We should aim to get all machines to the state where they can pass a run of the AQA_test_Pipeline job, but I suspect if we can eliminate the issues shown by the tests in those grinders (TARGET of testList TESTLIST=jdk_jdi_jdk8_0,jdk_jdi_jdk8_1,hotspot_jre_0,hotspot_jre_1,jdk_security3_0,hotspot_jre_0,hotspot_jre_1) we'll be most of the way there.

A number of the errors seem likely to be caused by the hostname not be resolvable i.e. ping $(hostname) doesn't work. I suspect we should look at switching the hostname on the machine to match the jenkins name, as the machines have entries for those in /etc/hosts

smlambert · 2024-01-08T21:38:38Z

From Deep History:

Machines with the Hostname and service name not provided issue:
test-osuosl-aix72-ppc64-3
test-osuosl-aix72-ppc64-4
test-osuosl-aix72-ppc64-5

Machines that do not have that issue:
test-osuosl-aix72-ppc64-1
test-osuosl-aix72-ppc64-2

sxa · 2024-01-11T14:10:39Z

Looking at one of them - adopt06 has this entry in /etc/hosts but has a hostname of adopt06 which is not resolvable (I would expect that the hostname was updated to match the inventory, but not the actual machine hostname, which is why we get such failures (TL;DR ping $(hostname) doesn't work. I'll run tests with the jdk_lang_0,jdk_jdi_jdk8_0,jdk_jdi_jdk8_1 targets to confirm:

Test on -3
Test on -6
Test on -6 after adding adopt06 to /etc/hosts (Passed green)

sxa · 2024-01-12T10:04:09Z

Noting that the /etc/hosts will get replaced by the regular refreshes from AWX.

I've kicked off the following after reinstating the originally deployed line in /etc/hosts with the adoptXX name while the one that ansible updates (external IP and hostname) is left as-is. This should ensure that the adoptXX line does not disappear. In the general case we have three options and we should decide which is the most appropriate:

What I've done as a temporary fix - re-add the adoptXX to /etc/hosts with the local network IP address - I've done this on test machines 03 through 05
Use a hostname that is in the osuosl DNS so it can be resolved that way e.g. p8-aix1-adopt04 (test machine -2) is set up this way and the hostname gets resolved by the DNS at 140.211.166.130. -1 is the same.
Change the hostname on the machine to match our inventory so that it is resolved via the line that ansible is adding to /etc/hosts. I've now done this on -6 so we can verify the outcome.

New test jobs (If Full sanity=N it means I'm just running the same three targets as earlier):

Grinder	Full sanity?	host	config	result
8457	N	test-6	3	PASS
8458	N	test-5	1	PASS
8459	Y	test-4	1	Failed [1]
8460	N	test-3	1	Failed [1]
8461	Y	test-2	2	PASS

[1] - Netowkr tests were all good so the problem in this issue is resolved, however these two runs had a failure in java/lang/Thread:

Execution failed: `main' threw exception: java.lang.OutOfMemoryError: unable to create new native thread

sxa · 2024-01-12T18:45:00Z

Noting that the AWX deploy is overwriting my new line, so for now I've added hosts_file to the list of roles to skip in the regular playbook deploymnet on AIX

sxa · 2024-01-14T10:21:46Z

Running 100 iteration with the failing thread test (java/lang/ThreadLocal/TestThreadId.java.TestThreadId) on two machines:

Machine	result
test-3	All passed
test-5	All passed

And ten instances of jdk_lang_0 on:

Machine	result
test-4	Mix of pass and fail

Suggests it's a load issue of some sort when running the whole suite, although the tests are using concurrency:1 but this probably needs tobe a separate issue as the original problem described in this issue is now resolved (although needs an improved playbook fix since I' vemanually patched /etc/hosts)

sxa · 2024-01-15T09:10:57Z

Intermittent testThreadId failure is covered under adoptium/aqa-tests#2189

sxa · 2024-01-15T09:17:33Z

As a follow-up to the proposals above, noting that for other operating systems:

We change the hostname in the UNIX playbooks to match the name from the inventory.yml file:

infrastructure/ansible/playbooks/AdoptOpenJDK_Unix_Playbook/roles/adopt_etc/tasks/main.yml

Line 25 in d2303c7

- name: Set hostname to inventory_hostname
On windows we have logic to do a hostname change, but it's only invoked if the machine name ends in a hypen character:

infrastructure/ansible/playbooks/AdoptOpenJDK_Windows_Playbook/roles/Common/tasks/main.yml

Line 41 in d2303c7

- name: Check hostname doesn't end in '-'

(From test-azure-win2019-x64-1: jdk_rmi java/rmi/Naming/legalRegistryNames/LegalRegistryNames.java test failure #2133)

On this basis it is likely that making a similar change to the UNIX playbook on the AIX machines is the preferred option here, however given the proximity to the January release cycle I suggest we pause anything more for now (although comments/discussion on the options are still welcome) We could also use /etc/motd to remind people on login what the adoptXX name is (Similar to what we do for the RISC-V machines at the PLCTlab.

sxa · 2024-01-15T10:58:51Z

Closing as 3344 has been split out to cover a permanent solkution going forward so we don't need to block the hosts_file rule.

smlambert changed the title ~~NoAddresses.java in jdk_security4_1 FAILED in Test_openjdk20_hs_sanity.openjdk_ppc64_aix with UnknownHostException~~ Many network tests failing on AIX in JDK20 (jdk_security4_1 FAILED in with UnknownHostException) Jul 17, 2023

This was referenced Jul 17, 2023

AIX 7.1 to reach end of support after April #3030

Closed

July 2023 Release JDK17 triage adoptium/aqa-tests#4677

Closed

smlambert changed the title ~~Many network tests failing on AIX in JDK20 (jdk_security4_1 FAILED in with UnknownHostException)~~ Many network tests failing on AIX in JDK20 (with UnknownHostException: Hostname and service name not provided) Sep 5, 2023

smlambert transferred this issue from adoptium/aqa-tests Sep 5, 2023

adamfarley mentioned this issue Sep 6, 2023

Weekly build triage for the week starting Saturday 2023/09/02 adoptium/temurin-build#3473

Closed

adamfarley mentioned this issue Oct 13, 2023

Weekly build triage for the week starting Saturday 2023/10/07 adoptium/temurin-build#3501

Closed

sxa removed this from Adoptium 3Q 2023 Plan Nov 2, 2023

sxa added this to Adoptium 4Q 2023 Plan Nov 2, 2023

sxa added this to the Backlog milestone Jan 3, 2024

sxa mentioned this issue Jan 3, 2024

Problem machines for release #2662

Open

smlambert added this to 2024 1Q Adoptium Plan Jan 8, 2024

smlambert removed this from Adoptium 4Q 2023 Plan Jan 8, 2024

smlambert moved this to Todo in 2024 1Q Adoptium Plan Jan 8, 2024

smlambert added the critical label Jan 8, 2024

smlambert mentioned this issue Jan 14, 2024

Jan CPU Dry-run Triage - Temurin JDK17 adoptium/aqa-tests#4960

Closed

sxa mentioned this issue Jan 15, 2024

Resolve AIX hostname resolution issues #3344

Open

sxa closed this as completed Jan 15, 2024

sxa self-assigned this Jan 15, 2024

github-project-automation bot moved this from Todo to Done in 2024 1Q Adoptium Plan Jan 15, 2024

sxa modified the milestones: Backlog, 2024-01 (January) Jan 15, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Many network tests failing on AIX in JDK20 (with UnknownHostException: Hostname and service name not provided) #3178

Many network tests failing on AIX in JDK20 (with UnknownHostException: Hostname and service name not provided) #3178

smlambert commented Jul 17, 2023 •

edited

Loading

smlambert commented Sep 5, 2023

smlambert commented Sep 5, 2023

aixtools commented Sep 5, 2023

smlambert commented Sep 5, 2023

smlambert commented Sep 5, 2023

adamfarley commented Sep 6, 2023

aixtools commented Sep 6, 2023

aixtools commented Sep 6, 2023

adamfarley commented Sep 6, 2023

adamfarley commented Sep 6, 2023 •

edited

Loading

aixtools commented Sep 6, 2023 via email

adamfarley commented Oct 13, 2023

sxa commented Nov 2, 2023 •

edited

Loading

smlambert commented Jan 8, 2024

sxa commented Jan 11, 2024 •

edited

Loading

sxa commented Jan 12, 2024 •

edited

Loading

sxa commented Jan 12, 2024

sxa commented Jan 14, 2024 •

edited

Loading

sxa commented Jan 15, 2024

sxa commented Jan 15, 2024 •

edited

Loading

sxa commented Jan 15, 2024

Many network tests failing on AIX in JDK20 (with UnknownHostException: Hostname and service name not provided) #3178

Many network tests failing on AIX in JDK20 (with UnknownHostException: Hostname and service name not provided) #3178

Comments

smlambert commented Jul 17, 2023 • edited Loading

smlambert commented Sep 5, 2023

smlambert commented Sep 5, 2023

aixtools commented Sep 5, 2023

smlambert commented Sep 5, 2023

smlambert commented Sep 5, 2023

adamfarley commented Sep 6, 2023

aixtools commented Sep 6, 2023

aixtools commented Sep 6, 2023

adamfarley commented Sep 6, 2023

adamfarley commented Sep 6, 2023 • edited Loading

aixtools commented Sep 6, 2023 via email

adamfarley commented Oct 13, 2023

sxa commented Nov 2, 2023 • edited Loading

smlambert commented Jan 8, 2024

sxa commented Jan 11, 2024 • edited Loading

sxa commented Jan 12, 2024 • edited Loading

sxa commented Jan 12, 2024

sxa commented Jan 14, 2024 • edited Loading

sxa commented Jan 15, 2024

sxa commented Jan 15, 2024 • edited Loading

sxa commented Jan 15, 2024

smlambert commented Jul 17, 2023 •

edited

Loading

adamfarley commented Sep 6, 2023 •

edited

Loading

sxa commented Nov 2, 2023 •

edited

Loading

sxa commented Jan 11, 2024 •

edited

Loading

sxa commented Jan 12, 2024 •

edited

Loading

sxa commented Jan 14, 2024 •

edited

Loading

sxa commented Jan 15, 2024 •

edited

Loading