Thanks Dan, Jeff for sharing so much of infomation. I will try the below option and share the result.
Try upgrading from 1.6.0_25 to 1.6.0_35. Thanks & Regards, Aditi On Thu, Sep 13, 2012 at 7:39 PM, Jeffrey Janner <jeffrey.jan...@polydyne.com > wrote: > > -----Original Message----- > > From: Daniel Mikusa [mailto:dmik...@vmware.com] > > Sent: Wednesday, September 12, 2012 2:14 PM > > To: Tomcat Users List > > Subject: Re: HTTP NIO connector not supporting IPv6 > > > > On Sep 12, 2012, at 2:44 PM, Jeffrey Janner wrote: > > > > >> -----Original Message----- > > >> From: Daniel Mikusa [mailto:dmik...@vmware.com] > > >> Sent: Wednesday, September 12, 2012 10:00 AM > > >> To: Tomcat Users List > > >> Subject: Re: HTTP NIO connector not supporting IPv6 > > >> > > >> On Sep 12, 2012, at 1:29 AM, Aditi Sinha wrote: > > >> > > >>> Thanks Dan, Jeff. > > >>> > > >>> > > >>> > > >>> There are no errors in catalina.log file. > > >>> > > >>> The connector tags are defined as below in server.xml. This > > >>> configuration does not support IPv6. > > >>> > > >>> > > >>> <Connector port="8080" protocol="*HTTP/1.1*" > > >> connectionTimeout="20000" > > >>> redirectPort="8443" /> > > >>> > > >>> > > >>> > > >>> <Connector port="8443" protocol="* > > >>> org.apache.coyote.http11.Http11NioProtocol*" SSLEnabled="true" > > >>> maxThreads="150" > > >>> > > >>> scheme="https" secure="true" clientAuth="false" > > >> sslProtocol="TLS" > > >>> keystoreFile="xxxxxx" keystorePass="xxxxxx"/> > > >>> > > >> > > >> Tried it on my MBP (10.7 w/Java 1.6.0_35) and it worked fine. Tried > > >> on Windows XP (only version I have available) w/ Java 1.6.0_35 and > > >> was able to replicate the problem behavior. According to the > > >> following bug report this is a limitation of the OS / JVM. Looks > > >> like a recent versions of Windows and a recent version of the JVM > > are > > >> required to resolve this. > > >> > > >> Try upgrading from 1.6.0_25 to 1.6.0_35. > > >> > > >> http://bugs.sun.com/view_bug.do?bug_id=6230761 > > >> > > >> Dan > > >> > > > One workaround is to explicitly define the IPv4 and IPv6 addressing > > in the <connector>. > > > That is, add address="0.0.0.0" for any IPv4 connectors and > > > address="[::]" for any IPv6 connectors. (Or use your real addresses > > instead of the "any" addresses listed here.) This means setting up 2 > > sets of connectors for each port/protocol, but there's nothing wrong > > with being explicit. > > > This is what I had to do to get the APR protocol to set up for IPv4. > > > > > > With all due respect, I do not think that this is going to work for the > > poster. You're saying that your workaround was for an issue with the > > APR connector, but the poster is using the NIO connector. The APR > > connector does not use NIO or the JVM, so I don't think your workaround > > is relevant. In fact I tried your workaround previously without > > success. The problem is that on older versions of Windows (pre-vista) > > and older versions of the JVM, the NIO libraries do not support IPv6 > > (see bug report). > > > > As a side note, the poster could switch from NIO to the APR connector, > > and it would likely resolve his problem (just like he reported > > switching to the BIO connector resolved his problem). Just assuming > > that the poster wants to stick with the NIO connector. > > > > Dan > > > > Yes, Dan. But the OP stated he is using Windows Server 2008, which is the > server version of "Vista". Since he doesn't say he's on 2008 R2 (which is > Windows 7 Server), so we'll assume that he's at least at a Vista-level > networking. > > The bug you mention is specifically about dual-mode socket support, that > is, being able to specify the port and not the address and having Java set > up to sockets which listen/talk on both the IPv4 & IPv6 addresses of the > machine. At least if I read it correctly. Re-reading it, it might not be > there for NIO, no matter what you try. My work-around would specifically > show whether the support is there or not for IPv6 in NIO, i.e., use it as a > diagnostic tool, if nothing else. > > But yes, reading the bug, really closely, makes it appear that the support > for NIO IPv6 was not back-ported to JDK6 until this last July, specifically > 6u33-b34. So the OP should probably start up upgrading his Java to > jdk1.6_34 or later and see if that doesn't fix it first. > > Personally, I prefer the APR to NIO, but mainly because SSL is easier to > manage under APR. I believe there are some esoteric advantages to using > NIO, but I'd have to go back to the comparison chart to tell you what they > are. The OP may have a specific use case that requires those features. > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org > For additional commands, e-mail: users-h...@tomcat.apache.org > >
