Bug 900793 (JBEWS-197)
| Summary: | JK status manager - mass nodes handling works strange | ||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [JBoss] JBoss Enterprise Web Server 2 | Reporter: | Jan Stefl <jstefl> | ||||||||||||||||||
| Component: | mod_jk | Assignee: | Weinan Li <weli> | ||||||||||||||||||
| Status: | CLOSED EOL | QA Contact: | Libor Fuka <lfuka> | ||||||||||||||||||
| Severity: | medium | Docs Contact: | |||||||||||||||||||
| Priority: | medium | ||||||||||||||||||||
| Version: | 2.0.0 | CC: | jclere, jdoyle, jstefl, lfuka, mturk, pcheung, pslavice, rsvoboda | ||||||||||||||||||
| Target Milestone: | --- | ||||||||||||||||||||
| Target Release: | 2.0.0 | ||||||||||||||||||||
| Hardware: | Unspecified | ||||||||||||||||||||
| OS: | Unspecified | ||||||||||||||||||||
| URL: | http://jira.jboss.org/jira/browse/JBEWS-197 | ||||||||||||||||||||
| Whiteboard: | |||||||||||||||||||||
| Fixed In Version: | Doc Type: | Bug Fix | |||||||||||||||||||
| Doc Text: |
Due to an issue with the mod_jk version shipped with the JBoss Enterprise Web Server, multiple nodes using status manager URL were not managed properly.
The updated version of mod_jk in JBoss Enterprise Web Server 2.0.1 uses the status manager to manage multiple nodes as expected.
|
Story Points: | --- | ||||||||||||||||||
| Clone Of: | Environment: | ||||||||||||||||||||
| Last Closed: | 2019-06-13 12:09:47 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: | |||||||||||||||||||||
| Attachments: |
|
||||||||||||||||||||
Attachment: Added: jk-after-mass-disabling.log Attachment: Added: jk-after-mass-disabling-correct-reponse.log Attachment: Added: jk-after-mass-disabling-incorrect-reponse.log Attachment: Added: jk-after-mass-enabling.log Attachment: Added: jk-after-httpd-start.log Attachment: Added: mod_jk.conf Attachment: Added: uriworkermap.properties Attachment: Added: workers.properties Link: Added: This issue relates to JBPAPP-9551 Resolving because this is a duplicate of JBPAPP-9315. Reassigning to Permaine for concurrence. The duplicate is for EAP, let's track EWS using this JIRA. Mladen, is this something that we'll wait for a fix upstream? Yes. Those kind of issues needs to be reported and fixed upstream first. Jan, please file a bug upstream so that it gets fixed upstream and we'll inherit the fix when a new version is released from the upstream community. thanks! Ok, we use mod_jk version *1.2.36*. Actual version is *1.2.37*. Is seems that *1.2.37* + patch http://svn.apache.org/viewvc?view=revision&revision=1354021 Should fix the problem. https://issues.apache.org/bugzilla/show_bug.cgi?id=53762#c1 If relevant build will be created, I will test it. Patch applied. Will be part of ER10 I find out that was error in a test and problem still appears. Is there any reason why we use mod_jk 1.2.36 since 1.2.37 is out? Affects: Added: Release Notes Can someone elaborate on how the problem manifests? I can't get a very good idea from the description what exactly the problem visible to the user/error is. Release Notes Docs Status: Added: Documented as Known Issue Writer: Added: mhusnain Release Notes Text: Added: Attempts to manage multiple nodes using the status manager URL such as the following does not work correctly: cmd=update&mime=prop&w=router&att=vwa&val0=1&val1=1 As a workaround for this problem, use the status manager's web interface instead. Release Notes Docs Status: Removed: Documented as Known Issue Writer: Removed: mhusnain Release Notes Text: Removed: Attempts to manage multiple nodes using the status manager URL such as the following does not work correctly: cmd=update&mime=prop&w=router&att=vwa&val0=1&val1=1 As a workaround for this problem, use the status manager's web interface instead. Docs QE Status: Removed: NEW Assigning to Weinan to track if this fix gets in future releases of EWS. upgrading mod_jk should fix the problem. Not fixed in EWS 2.0.1 ER1 on Solaris build. Windows and Linux ER1 seems ok. But the code is the same. And it the same as for EWS 2.0.0 although we have 1.1.26 + patches which equals to 1.1.27. Sorry I meant 1.2.36 and 1.2.37 There were no code changes between EWS 2.0.0 and 2.0.1 Updated version did not add any new code because it consists of all the patches we already have for 1.2.36 Patch http://svn.apache.org/viewvc?view=revision&revision=1354021 was not applied in Solaris build. (In reply to comment #20) > Patch http://svn.apache.org/viewvc?view=revision&revision=1354021 was not > applied in Solaris build. Think it was, see: http://pkgs.devel.redhat.com/cgit/rpms/mod_jk/commit/?h=jb-ep-6-xb&id=63df3b40997ef53be5a39baea5d8af501ad1384f It seems in sources that patch was not applied also in Linux, Windows jk_status.c. How is it possible patch is not seen in build sources ? it should be mod_jk 1.2.37 + patch, see comm 8 Patch applied to RHEL versions: mod_jk-1.2.37-2.redhat_1.ep6.el5 - https://brewweb.devel.redhat.com/buildinfo?buildID=270458 mod_jk-1.2.37-2.redhat_1.ep6.el6 - https://brewweb.devel.redhat.com/buildinfo?buildID=270457 Will be included in ER2 Added DocText. @Wei Nan Li, can you please review the Doc Text content? Hi Mandar, there is a typo in text: 'Jboss' should be 'JBoss' Seems to be fixed only for windows. Still not fixed for RHEL and Solaris (for Solaris even verified that it doesn't contain the patch via src zip, for RHEL the src zip is broken) The Solaris zip is broken, after the zips updated it should be fine. For RHEL, I've checked the RHEL5 mod_jk and seems the patch is applied correct. RHEL6 is broken because the zip is not correct. Radim, could you please explain what's the problem on RHEL5? It's the patch not applied or the patch not work as expected? Actually I have tried it only with RHEL6 x86_64. As the zip archive with sources is common for both RHEL5 and RHEL6, I have anticipated that it will be thus broken for all RHEL platforms. VERIFIED on EWS-2.0.1-ER2 on RHEL5, RHEL6 on i386, x86_64 Zip fixed, return back to QE (according to Michal's previous comment, I propose this one set to VERIFIED). VERIFIED on EWS-2.0.1-ER2 on Solaris 10,11 (i386, x86_64, SPARC) and MS Windows 2008 (i686, x86_64), 2008R2 (x86_64) Excluded from Release Notes as part of the internal group. |
Affects: Release Notes project_key: JBEWS I have one HTTPD server balanced 2 Tomcats. I try to disable all Tomcats nodes at once smth. like: {code} curl -G -d "cmd=update&mime=prop&w=router&att=vwa&val0=1&val1=1" localhost/jkmanager/ {code} And expect returned HTTPD code 503 for multiple requests. In web browsers (Chrome, Firefox) it is behave strange. During the refreshes I sometimes see served response from Tomcat, but I would suppose, that could be same browser problem. This strange (random) behaviour is valid even for running wget muliple times. {code} wget --no-cache -O - 127.0.0.1/testapp {code} After that I want to start all nodes, but it is not working. Nothing happend and behaviour remains same as I described above. {code} curl -G -d "cmd=update&mime=prop&w=router&att=vwa&val0=0&val1=0" localhost/jkmanager/ {code} Btw. shared memory file (JkShmFile) is on local disk (no NFS). See my configurations and logs in the attached files. This JIRA is duplicate for JBPAPP-9315