Bug 900978 (JBEWS-85)
Summary: | IPv6 :: address slowing down Tomcat6/7 on MS Windows | ||
---|---|---|---|
Product: | [JBoss] JBoss Enterprise Web Server 2 | Reporter: | Libor Fuka <lfuka> |
Component: | tomcat6, tomcat7 | Assignee: | Coty Sutherland <csutherl> |
Status: | CLOSED DEFERRED | QA Contact: | Michal Karm Babacek <mbabacek> |
Severity: | medium | Docs Contact: | |
Priority: | medium | ||
Version: | 2.0.0 | CC: | jclere, mturk, pcheung, pslavice, rsvoboda, smumford, weli |
Target Milestone: | --- | ||
Target Release: | TBD EWS | ||
Hardware: | Unspecified | ||
OS: | Unspecified | ||
URL: | http://jira.jboss.org/jira/browse/JBEWS-85 | ||
Whiteboard: | |||
Fixed In Version: | Doc Type: | Known Issue | |
Doc Text: |
In JBoss Enterprise Web Server, when an IPv6 address is used to start Tomcat 6 or 7 on Windows and then the relevant version of Tomcat is shut down, the shutdown process is unexpectedly slow. Additionaly, during the shutdown process, the <filename>catalina.log</filename> file logs the following warning:
<screen>WARNING: Acceptor thread [http-0%3A0%3A0%3A0%3A0%3A0%3A0%3A0-8080-Acceptor-0] failed to unlock. Forcing hard socket shutdown</screen>
This is a known issue for JBoss Enterprise Web Server 3.0 and currently no workaround is available.
|
Story Points: | --- |
Clone Of: | Environment: | ||
Last Closed: | 2017-08-09 12:57:30 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: | |||
Deadline: | 2012-10-05 |
Description
Libor Fuka
2012-10-02 15:35:06 UTC
Link: Added: This issue relates to JBPAPP-9551 I don't know. It's a Tomcat problem. Ask David if he applied the mentioned upstream fix Assigning to David, since it's Tomcat/java problem $ tail -f logs/catalina.2012-10-05.log Oct 5, 2012 12:14:48 PM org.apache.coyote.ajp.AjpAprProtocol pause INFO: Pausing Coyote AJP/1.3 on ajp-8009 Oct 5, 2012 12:14:49 PM org.apache.catalina.core.StandardService stop INFO: Stopping service Catalina Oct 5, 2012 12:14:49 PM org.apache.coyote.http11.Http11AprProtocol destroy INFO: Stopping Coyote HTTP/1.1 on http-0%3A0%3A0%3A0%3A0%3A0%3A0%3A0-8080 Oct 5, 2012 12:15:00 PM org.apache.tomcat.util.net.AprEndpoint stop WARNING: Acceptor thread [http-0%3A0%3A0%3A0%3A0%3A0%3A0%3A0-8080-Acceptor-0] failed to unlock. Forcing hard socket shutdown. Oct 5, 2012 12:15:00 PM org.apache.coyote.ajp.AjpAprProtocol destroy INFO: Stopping Coyote AJP/1.3 on ajp-8009 Well that is 10 seconds... Not a big deal. Yes, Its not so big problem. It is related to AS7-3834 in fact... The unlockAccept() will used the "::" address to connect to server to unlock the accept that won't work. That is a upstream bug... Reported as upstream bugzilla 53971 Should be assigned for inclusion in future EWS release. Think we should just check in shutdownAccept if the string address is "::" and try to connect with "::1" I'm just shooting at random here :) Assigning to David for inclusion in future EWS release. Please file a doc bz and file this into Known issues section. I will, its not resolved. Release Notes Docs Status: Added: Not Yet Documented Release Notes Docs Status: Removed: Not Yet Documented Added: Documented as Known Issue Release Notes Text: Added: When an IPv6 address is used to start Tomcat 6 or 7 on Windows (for example <Connector port="8080" protocol="HTTP/1.1" address="::" />) and results in a significantly slowed down shutdown process when Tomcat is sunsequently shut down. During the shutdown process, the catalina.log file logs a warning for this shutdown: WARNING: Acceptor thread [http-0%3A0%3A0%3A0%3A0%3A0%3A0%3A0-8080-Acceptor-0] failed to unlock. Forcing hard socket shutdown Affects: Added: Release Notes Writer: Added: mhusnain Release Notes Docs Status: Removed: Documented as Known Issue Writer: Removed: mhusnain Release Notes Text: Removed: When an IPv6 address is used to start Tomcat 6 or 7 on Windows (for example <Connector port="8080" protocol="HTTP/1.1" address="::" />) and results in a significantly slowed down shutdown process when Tomcat is sunsequently shut down. During the shutdown process, the catalina.log file logs a warning for this shutdown: WARNING: Acceptor thread [http-0%3A0%3A0%3A0%3A0%3A0%3A0%3A0-8080-Acceptor-0] failed to unlock. Forcing hard socket shutdown Docs QE Status: Removed: NEW I can do that, anyone against to be the public bug ? It is reported upstream so it MUST be public. Set to public Thanks, Libor. Assuming there are no objections to doc text and setting flag to + to indicate a final draft for release notes. There are several upstream and jira threads some with several patches that affect related bits of code. The code involved has been refactored in 7.0.40 and again in 7.0.47. This bug was first recorded in 2012. We've since updated the source. Maybe this bz isn't relevant any more and should be updated to modified so it can be verified. Problem still persists in T 6.0.41 and T 7.0.54. Stopping is lasting around 10-15s with address="::" more than without it. and warning in catalina.out (T6 and T7) is still the same: Jun 27, 2014 5:08:19 AM org.apache.tomcat.util.net.AprEndpoint stopInternal WARNING: Acceptor thread [http-apr-0:0:0:0:0:0:0:0-8080-Acceptor-0] failed to unlock. Forcing hard socket shutdown. Is it still the same with the BIO/JIO connector? Its only problem when we use the APR/native connector. Keep it as a known issue at least for apr/native. The work-around now is to use BIO/JIO or NIO (tomcat7), it was just partially fixed. |