Bug 1552040
| Summary: | Sometimes Wildfly 12 process does not exit on RHEL 6 (32/64 bit) and IBM JDK 8 | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Miroslav Novak <mnovak> | ||||||
| Component: | java-1.8.0-ibm | Assignee: | jiri vanek <jvanek> | ||||||
| Status: | CLOSED INSUFFICIENT_DATA | QA Contact: | BaseOS QE - Apps <qe-baseos-apps> | ||||||
| Severity: | high | Docs Contact: | |||||||
| Priority: | high | ||||||||
| Version: | 6.9 | CC: | bugproxy, hannsj_uhl, jjarvis, jkachuck, mnovak | ||||||
| Target Milestone: | rc | ||||||||
| Target Release: | 6.10 | ||||||||
| Hardware: | All | ||||||||
| OS: | Linux | ||||||||
| Whiteboard: | |||||||||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |||||||
| Doc Text: | Story Points: | --- | |||||||
| Clone Of: | Environment: | ||||||||
| Last Closed: | 2018-04-05 07:03:54 UTC | Type: | Bug | ||||||
| Regression: | --- | Mount Type: | --- | ||||||
| Documentation: | --- | CRM: | |||||||
| Verified Versions: | Category: | --- | |||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||
| Embargoed: | |||||||||
| Bug Depends On: | |||||||||
| Bug Blocks: | 1414846, 1425546 | ||||||||
| Attachments: |
|
||||||||
|
Description
Miroslav Novak
2018-03-06 11:54:47 UTC
------- Comment From chavez.com 2018-03-06 16:39 EDT------- Raised problem ticket with Java L3 containing current description. ------- Comment From chavez.com 2018-03-07 08:01 EDT------- Response from Java L3: I have been trying to reproduce the hang - but not successful so far. The application shuts down properly everytime. From the native threads - my assumption is that the issue may turn out to be something to do with the OS. Almost all the native threads have the following stacktop: #0 0x00bd4424 in __kernel_vsyscall () #1 0x00b2243c in pthread_cond_wait@@GLIBC_2.3.2 () from /lib/libpthread.so.0 Perhaps the OS is not bringing these threads out of the cond wait. The Java VM has already progressed to shut down - and the signal handler services maybe already off - because of which the logs are not generated. However, coredump from the hang will help us to find more details. Could you please check if you can generate a coredump using gcore? You may follow these steps: 1. Keep the ulimits set to unlimited: # ulimit -c unlimited # ulimit -f unlimited 2. Whenever you see the hang during the shutdown, use gcore on the JVM pid: # gcore <java_pid> Hopefully this will generate a coredump from the hang. 3. Jextract the coredump: [path to JDK]/jre/bin/jextract [coredump] This should create a zip file - containing the coredump and the related libraries. Please upload this for us. Thanks. ------- Comment From chavez.com 2018-03-20 11:51 EDT------- Any update? Java L3 is considering closing the problem ticket on March 23 if nothing is received by then. Thanks. Sorry guys for late response. I was assigned to another task. I understand that this might issue on RedHat site as well. I took the whole day of start/stop WF12 until issue could be hit again. I also figured out that somehow our testing framework catched kill -3 signal so I needed to do it without it. I'm attaching java core dump. I did not manage to take a look at it. Created attachment 1410618 [details]
javacore dump
------- Comment From chavez.com 2018-03-20 13:18 EDT------- Understand but do you have an actual core dump (not just the javacore text file) that you can run jextract on so it can create an archive with the core and all the binaries and libraries necessary to analyze it as previously requested? Let me know. Thanks. Hello IBM\Miroslav Please note at this point in the RHEL 6.10 release. Without an exception or blocker request this will not be able to make RHEL 6.10. If this will be requested for this BZ. Please confirm what a client would see in the field? What it would mean if this was not fixed in RHEL 6.10. Miroslav, Please supply the logs ASAP if this is required to be fixed in RHEL 6.10. Thank You Joe Kachuck Sorry guys, I've kept it running in cycles for last 4 days but still was not successful to reproduce it again. I'll keep it spinning on more machines to increase chances to hit it. Thanks Joseph for update. I understand that we'll not meet RHEL 6.10 window if this is not investigated in time. I've kept the test spinning on multiple machines (IBM JDK 8...10, RHEL 6 32/64bit) the whole week but could not reproduce the issue. There were thousands of runs. It might be that issue is gone or it was some env issue. I would suggest to close this bz/ticket for now and re-open if this occur again. ------- Comment From chavez.com 2018-04-04 13:21 EDT------- OK. Thanks for the update. Closing. |