Re: Tomcat 6.0.29 using more and more RAM until it collapses?
On 16/11/2010 05:08, Brian wrote: I still have to find the reason of the leak. Tomcat attempts to find notify you of potential memory leaks. You reported messages in your logs at app reload. Did you tell us what they are yet? p 0x62590808.asc Description: application/pgp-keys signature.asc Description: OpenPGP digital signature
Re: Tomcat Going down Frequently
On 16/11/2010 07:34, Amol Puglia wrote: Hello Pid, Thanks for the updates.It would be great if you could let us know whether the issue is in application server or application. Well, the problem class isn't in an org.apache package, it's in a com.ericsson package. Is that your code? LoginBean.java, line 2366. There's a lot of code in that file, but you could check to see what's happening at that line. In the a line above in that, in the stacktrace, there's a call to StandardSessionFacade.getAttributeNames which I'm not sure about, I'd expect to see a call to HttpSession rather than a call to an internal component of Tomcat. I think your code is attempting to flush the session clean, without checking that the session is available to do so (which it isn't). p Also we are using apache to load balance tomcat instances using following modules. LoadModule proxy_module modules/mod_proxy.so LoadModule proxy_connect_module modules/mod_proxy_connect.so LoadModule proxy_ftp_module modules/mod_proxy_ftp.so LoadModule proxy_http_module modules/mod_proxy_http.so LoadModule proxy_scgi_module modules/mod_proxy_scgi.so LoadModule proxy_ajp_module modules/mod_proxy_ajp.so LoadModule proxy_balancer_module modules/mod_proxy_balancer.so Is there anything we have to tune configuration on apache,tomcat or application,please suggest. --- On Mon, 11/15/10, Pid p...@pidster.com wrote: From: Pid p...@pidster.com Subject: Re: Tomcat Going down Frequently To: Tomcat Users List users@tomcat.apache.org Date: Monday, November 15, 2010, 10:04 PM On 15/11/2010 11:03, Amol Puglia wrote: at org.apache.catalina.session.StandardSessionFacade.getAttributeNames(StandardSessionFacade.java:120) at com.ericsson.mars.jspbean.LoginBean.logout(LoginBean.java:2366) at com.ericsson.mars.jspbean.LoginBean.valueUnbound(LoginBean.java:2450) The problem is in your HttpSessionBindingListener LoginBean, it appears to be operating on the session object after the session has been invalidated. I don't know whether this could take Tomcat down, unless you've included* a System.exit call in a catch - which seems to be popular again all of a sudden. p * which is A Really Bad Idea. 0x62590808.asc Description: application/pgp-keys signature.asc Description: OpenPGP digital signature
Re: Tomcat Going down Frequently
Amol Puglia wrote: Hello Pid, Thanks for the updates.It would be great if you could let us know whether the issue is in application server or application. Also we are using apache to load balance tomcat instances using following modules. You most probably are /not/ using all of the following modules. Just having them in your configuration does not mean that you are actually using them. If you have them in your configuration, and are not using them, then you are probably wasting quite a bit of memory however. LoadModule proxy_module modules/mod_proxy.so LoadModule proxy_connect_module modules/mod_proxy_connect.so LoadModule proxy_ftp_module modules/mod_proxy_ftp.so LoadModule proxy_http_module modules/mod_proxy_http.so LoadModule proxy_scgi_module modules/mod_proxy_scgi.so LoadModule proxy_ajp_module modules/mod_proxy_ajp.so LoadModule proxy_balancer_module modules/mod_proxy_balancer.so Is there anything we have to tune configuration on apache,tomcat or application,please suggest. On the face of it, and considering that you are using mod_jk, my first recommendation as to tuning would be to comment out all of the above. You probably are not using any of them. As others have already commented, your problem seems definitely to be in the tomcat application, not in tomcat nor apache httpd. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
RE: the tomcat encoding
Hi Xu, Encoding is/can be a real pain. You have to ensure it is consistent throughout the web application to stand any chance. Start with the server.xml with the following declaration:- Connector port=80 redirectPort=443 URIEncoding=UTF-8 useBodyEncodingForURI=true / I also use a filter to set the following:- request.setCharacterEncoding(UTF-8); response.setContentType(text/html; charset=UTF-8); response.setCharacterEncoding(UTF-8); Within the final html output I always include the following meta tag:- meta http-equiv=Content-Type content=text/xhtml; charset=utf-8/meta Not sure all steps are required but they work for me. Of course if you use a database this also needs to support UTF-8 or your chosen encoding. Hope this helps Rob -Original Message- From: xu cheng [mailto:xcheng@gmail.com] Sent: 16 November 2010 05:59 To: users@tomcat.apache.org Subject: the tomcat encoding hi all: I wrote an app and there are some chinese content, and there are some code like this str = new String(chineseContentString.getBytes(iso-8859-1), utf-8); the app runs pretty well in the tomcat of my pc. however ,when it was deployed in other machine. all the Chinese content becomes messy code. so ,does anyone know how to got to encoding of the tomcat. or it was just iso-8859-1 and cannot be changed? thanks by the way , the edition is 6.0.20 - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: After manager says that there was a leak, how to use a profiler?
On 16/11/2010 04:12, Brian wrote: Maybe my question sounded too vague and repetitive. What I meant is something like this Is the a class loader (or something like that) than I should start analyzing? I mean, is there a route I should follow in order to see which objects are stucked? If the Tomcat Manager knows that something got stucked in the memory, how does it know that? Is there a clue we should start?. http://markmail.org/message/fcbvwapt6afyndxn Mark - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: the tomcat encoding
On 16/11/2010 05:59, xu cheng wrote: so ,does anyone know how to got to encoding of the tomcat. or it was just iso-8859-1 and cannot be changed? http://wiki.apache.org/tomcat/FAQ/CharacterEncoding Mark - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat Going down Frequently
Hello Andre, Thanks for the updates. I am not using mod_jk , I am using proxy modules for redirecting request from apache to tomcat. As you said that there is problem in application code, Is there anything we can point out where exactly is the problem in the code or configuration? --- On Tue, 11/16/10, André Warnier a...@ice-sa.com wrote: From: André Warnier a...@ice-sa.com Subject: Re: Tomcat Going down Frequently To: Tomcat Users List users@tomcat.apache.org Date: Tuesday, November 16, 2010, 2:54 PM Amol Puglia wrote: Hello Pid, Thanks for the updates.It would be great if you could let us know whether the issue is in application server or application. Also we are using apache to load balance tomcat instances using following modules. You most probably are /not/ using all of the following modules. Just having them in your configuration does not mean that you are actually using them. If you have them in your configuration, and are not using them, then you are probably wasting quite a bit of memory however. LoadModule proxy_module modules/mod_proxy.so LoadModule proxy_connect_module modules/mod_proxy_connect.so LoadModule proxy_ftp_module modules/mod_proxy_ftp.so LoadModule proxy_http_module modules/mod_proxy_http.so LoadModule proxy_scgi_module modules/mod_proxy_scgi.so LoadModule proxy_ajp_module modules/mod_proxy_ajp.so LoadModule proxy_balancer_module modules/mod_proxy_balancer.so Is there anything we have to tune configuration on apache,tomcat or application,please suggest. On the face of it, and considering that you are using mod_jk, my first recommendation as to tuning would be to comment out all of the above. You probably are not using any of them. As others have already commented, your problem seems definitely to be in the tomcat application, not in tomcat nor apache httpd. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
RE: Shutting down one instance of tomcat 6 from a listener
It works just fine doing it from the ServletContextListener. Thx a lot. But I was wondering how to do it outside of the application itself? Patrick Sauts. -Original Message- From: Pid [mailto:p...@pidster.com] Sent: lundi 15 novembre 2010 12:44 To: Tomcat Users List Subject: Re: Shutting down one instance of tomcat 6 from a listener On 15/11/2010 09:37, Patrick Sauts wrote: So, how to shut down properly the current application from a 'listener' leaving the other webapps deployed into Tomcat to continue their business? It depends on what the problem actually is, and when it occurs. If you do all of your initialisation in a ServletContextListener and throw an exception on fail, then the context won't start up. This is not the same as shutting it down after a certain period of time and after a fault condition occurs. Where is the code which is going to shut down the application? Presumably outside of the application itself? p - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat Going down Frequently
Amol Puglia wrote: Hello Andre, Thanks for the updates. I am not using mod_jk , I am using proxy modules for redirecting request from apache to tomcat. Ok, sorry, I confused your post with another one. But anyway, your configuration shows a whole lot of proxy modules, many of which do not seem to have any relation with proxying calls to Tomcat. Such as these : LoadModule proxy_connect_module modules/mod_proxy_connect.so LoadModule proxy_ftp_module modules/mod_proxy_ftp.so LoadModule proxy_scgi_module modules/mod_proxy_scgi.so The following /may/ have a relation with the proxying to Tomcat, but which one are you really using ? LoadModule proxy_http_module modules/mod_proxy_http.so LoadModule proxy_ajp_module modules/mod_proxy_ajp.so Usually, people use *either* mod_proxy_http *or* mod_proxy_ajp, so which one is being used ? Can you provide some information about the apache httpd configuration that shows what is being proxied and how ? As you said that there is problem in application code, Is there anything we can point out where exactly is the problem in the code or configuration? I thought that Pid already pointed that out. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Shutting down one instance of tomcat 6 from a listener
On 16/11/2010 10:46, Patrick Sauts wrote: It works just fine doing it from the ServletContextListener. Thx a lot. But I was wondering how to do it outside of the application itself? Assuming you can tell if the app is in a fault condition, you can use JMX to start, stop reload apps. p Patrick Sauts. -Original Message- From: Pid [mailto:p...@pidster.com] Sent: lundi 15 novembre 2010 12:44 To: Tomcat Users List Subject: Re: Shutting down one instance of tomcat 6 from a listener On 15/11/2010 09:37, Patrick Sauts wrote: So, how to shut down properly the current application from a 'listener' leaving the other webapps deployed into Tomcat to continue their business? It depends on what the problem actually is, and when it occurs. If you do all of your initialisation in a ServletContextListener and throw an exception on fail, then the context won't start up. This is not the same as shutting it down after a certain period of time and after a fault condition occurs. Where is the code which is going to shut down the application? Presumably outside of the application itself? p - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org 0x62590808.asc Description: application/pgp-keys signature.asc Description: OpenPGP digital signature
Re: Client not able with perform client-cert authentication with Tomcat 6.0.29 on APR
Hi Mark,
Thank you for the settings. I am not sure what is the APR/native connector
version, I am using the default APR/native connector in 6.0.29 (I do not
set/change APR on my Windows machine).
I am not sure why the client certificate authentication failed when my
client certificate was signed with SHA256 but client certificate
authentication worked perfectly when client certificate was signed with
SHA1.
From http://old.nabble.com/SHA256-digest-windows-0.9.8k--td26123008.html, it
mentioned developer required to include a call to
OpenSSL_add_all_algorithms()
instead calling SSL_library_init() which only adds the more commonly uses
SSL
algorithms. I am not sure where should I include this.
Can you advice how to solve my problem?
My APR connector settings:
Connector port=8443
protocol=org.apache.coyote.http11.Http11AprProtocol
SSLEnabled=true
maxThreads=150
scheme=https
secure=true
SSLCertificateFile=C:\usr\tomcat\tomcat.crt
SSLCertificateKeyFile=C:\usr\tomcat\tomcat.key
SSLVerifyClient=optional
SSLVerifyDepth=1
SSLCipherSuite=AES128-SHA:DES-CBC3-SHA
SSLCACertificateFile=C:\usr-files\client-cert-ca.crt /
On 13 November 2010 00:38, Mark Thomas ma...@apache.org wrote:
On 12/11/2010 16:27, Goo Sam Kong wrote:
Hi
I am running Tomcat 6.0.29 with JDK 1.6.0_22 on Windows XP.
APR/native connector version? SSL re-negotiation wasn't supported until
recently and the CVE-2009-3555 fixes further complicate things.
Connector SSLCACertificateFile=C:\usr-files\client-cert-ca.crt
SSLCertificateFile=C:\usr\tomcat\tomcat.crt
SSLCertificateKeyFile=C:\usr\tomcat\tomcat.key
SSLCipherSuite=AES128-SHA:DES-CBC3-SHA
SSLEnabled=true
SSLEngine=on
SSLVerifyClient=optional
maxThreads=150
port=8443
protocol=HTTP/1.1
scheme=https
secure=true
sslProtocol=TLS /
Is SSLEngine a valid attribute here? I don't see it in the Connector docs.
SSLVerifyClient=optional can (should?) be removed.
Is that SSLCipherSuite compatible with your client? Try removing that
setting until everything else is working.
The following settings are known to work:
Connector
port=8443
protocol=org.apache.coyote.http11.Http11AprProtocol
SSLEnabled=true
maxThreads=150
scheme=https
secure=true
SSLCertificateFile=${catalina.base}/conf/tomcathost-cert.pem
SSLCertificateKeyFile=${catalina.base}/conf/tomcathost-key.pem
SSLCACertificateFile=${catalina.base}/conf/cacert.pem /
Mark
-
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
RE: After manager says that there was a leak, how to use a profiler?
Hi Mark, I read your information, which I will quote here: 1. Find an app that you can't reload without OOME 2. Get a profiler - I like Yourkit Full disclosure: they give ASF committers a free copy 3. Reload you app once 4. Use the profiler to look for instances of WebappClassLoader 5. Look for the one with the started attribute == false 6. Trace the GC roots for this instance 7. Fix whatever is hanging on to references to the instance - there should not be any Well, I'm using Yourkit as well. I reloaded the app twice, and I found 4 instances of WebappClassLoader. Two of them correspond to my two running apps, which is correct, and the other two correspond to the leaked apps. Both of them have an instance that have the started attribute with a false. Great, so I guess I'm in the right path to find the problem. Now I'm inspecting one of these objects with the started=false. If I choose the report of Paths from GC roots from it, I see LOTS of items! And all of them have the loader of indication. All the items are classes programmed by me, so I guess I'm guilty, and not the people that programmed Tomcat ;-) . They are DTOs, facades, Struts actionforms, etc. Why are all these classes in the list? What should I check about them? Why are they still loaded? I guess there is something missing in my programming. Maybe I should do something about all the objects that are still alive that correspond to all the loaded classes? Maybe I need to do something in the contextDestroyed() method in the ServletContextListener in my app, in order to destroy all the objects properly? Thanks a lot for your help! Brian -Original Message- From: Mark Thomas [mailto:ma...@apache.org] Sent: Tuesday, November 16, 2010 04:37 AM To: Tomcat Users List Subject: Re: After manager says that there was a leak, how to use a profiler? On 16/11/2010 04:12, Brian wrote: Maybe my question sounded too vague and repetitive. What I meant is something like this Is the a class loader (or something like that) than I should start analyzing? I mean, is there a route I should follow in order to see which objects are stucked? If the Tomcat Manager knows that something got stucked in the memory, how does it know that? Is there a clue we should start?. http://markmail.org/message/fcbvwapt6afyndxn Mark - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
RE: Shutting down one instance of tomcat 6 from a listener
-Original Message- From: Pid [mailto:p...@pidster.com] Sent: mardi 16 novembre 2010 13:34 To: Tomcat Users List Subject: Re: Shutting down one instance of tomcat 6 from a listener On 16/11/2010 10:46, Patrick Sauts wrote: It works just fine doing it from the ServletContextListener. Thx a lot. But I was wondering how to do it outside of the application itself? Assuming you can tell if the app is in a fault condition, you can use JMX to start, stop reload apps. Of course, thank you. p Patrick Sauts. -Original Message- From: Pid [mailto:p...@pidster.com] Sent: lundi 15 novembre 2010 12:44 To: Tomcat Users List Subject: Re: Shutting down one instance of tomcat 6 from a listener On 15/11/2010 09:37, Patrick Sauts wrote: So, how to shut down properly the current application from a 'listener' leaving the other webapps deployed into Tomcat to continue their business? It depends on what the problem actually is, and when it occurs. If you do all of your initialisation in a ServletContextListener and throw an exception on fail, then the context won't start up. This is not the same as shutting it down after a certain period of time and after a fault condition occurs. Where is the code which is going to shut down the application? Presumably outside of the application itself? p - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Tomcat 6.0.29 Cookie.getMaxAge() always returns -1 regardless of actual value of maximum age
I run Tomcat 6.0.29 on Windows XP and IE 8.0.6001.18702. I noticed that getMaxAge() method of any cookie always returns -1 even for persistent cookie with real positive value of maximum age. IE treats this cookie correctly: cookie persists for amount of seconds specified by setMaxAge method. But getMaxAge always return -1. Based on specification getMaxAge must return actual value of maximum age of a given cookie, not default value. Is this a bug?
Re: Usage of Tomcat SSL in client-side class
Moley Harey wrote: Hello folks, I have a set of web services installed in Tomcat 6.0.26 that I would like to access using HTTPs, I have following the instructions in Tomcat web side to set up JSSE support and now I can access the web services using SSH. My question is what do I have to do in my Java client-side classes to work with HTTPs requests from Tomcat? Do I have to accept somehow the certificate provided by Tomcat in my Java classes? Well, I'm totally newbie in these security issues :-/ Only because you mention that you are a total newbie, and because I am myself not far above that level, some beginning pointers : It helps to give things their correct name (or acronym), so as to be able to find the correct documentation and understand their relationship. SSL is Secure Socket Layer. (Google : SSL RFC) SSH is an acronym for Secure Shell (Google : SSH RFC) HTTP is an acronym for HyperText Transport Protocol. (Google : HTTP RFC) HTTPS (with a big S) is an acronym for HTTP over SSL. (Google : HTTPS RFC) So you are not accessing the web services over SSH, you are accessing your web services over HTTPS, which itself lives above an SSL connection, which itself lives above a TCP connection, which itself lives above an IP link (which itself lives over an Ethernet link (probably) etc..). And on top of it all, you have the application (both at the client side, and at the server side). Schematically : Client applicationServer application Client container (?) (e.g. a JVM) Server container (e.g. JVM + Tomcat) Client O.S. Server OS Client SSL layer (optional) Server SSL layer (optional) Client OS TCP/IP layerServer OS TCP/IP layer Client-side Ethernet layerServer-side Ethernet layer Client-side hardware layer -- Server-side hardware layer When the client side application sends something to the server-side application, whatever data is sent goes through all the layers on the client side, then over the physical connection, then through all the layers on the server side up to the server application. The encryption/decryption, if any, takes place at the SSL layer level, automatically. So basically, to answer our question, you do not /have/ to do something special at the application level. The container and the OS take care of the interface, so that your application can just use the standard read and write and open and close and request.getParameters() calls, no matter if the connection uses an SSL layer or not. But you /can/ make use of some of the SSL-specific things, if you want. For example, the server-side application setup /could/ tell the container that it will only accept requests that come over a secure channel, and only if they come from a client identified as X or Y. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat Going down Frequently
Ok, so now we know roughly what your configuration is like, and there is not something obviously wrong with it (to me). Now the question is back as follows : In WARNING: Exception processing manager org.apache.catalina.session.standardmana...@a20981 background process java.lang.IllegalStateException: getAttributeNames: Session already invalidated at org.apache.catalina.session.StandardSession.getAttributeNames(StandardSession.java:1052) at org.apache.catalina.session.StandardSessionFacade.getAttributeNames(StandardSessionFacade.java:120) at com.ericsson.mars.jspbean.LoginBean.logout(LoginBean.java:2366) at com.ericsson.mars.jspbean.LoginBean.valueUnbound(LoginBean.java:2450) ... The above is a WARNING message, indicating that Tomcat has caught something unusual. The something unusual being a java class (LoginBean.java:2366) asking for a list of attributes of a session which is already invalidated (aka, finished). That's an application problem. But this should not by itself cause the whole Tomcat to go down. What follows however : Nov 12, 2010 1:33:09 PM org.apache.coyote.http11.Http11AprProtocol pause INFO: Pausing Coyote HTTP/1.1 on http-9083 Nov 12, 2010 1:33:09 PM org.apache.coyote.ajp.AjpAprProtocol pause INFO: Pausing Coyote AJP/1.3 on ajp-9011 Nov 12, 2010 1:33:10 PM org.apache.catalina.core.StandardService stop INFO: Stopping service Catalina Nov 12, 2010 1:33:11 PM org.apache.coyote.http11.Http11AprProtocol destroy INFO: Stopping Coyote HTTP/1.1 on http-9083 Nov 12, 2010 1:33:11 PM org.apache.coyote.ajp.AjpAprProtocol destroy INFO: Stopping Coyote AJP/1.3 on ajp-9011 are mere INFO-level messages from Tomcat, showing basically that Tomcat is stopping, nicely and quietly, as if it had been asked politely to stop. It is not going down in flames, it is just cleaning up and stopping nicely, as requested. So what is asking it to stop ? Note also that as per the logfile, this orderly stop happens at 1:33:11, which is almost *4 minutes* later than the original WARNING message dated 1:29:23. That is quite an eternity, even by Java standards. So I am not even sure that the two things are related. Amol Puglia wrote: Hello Andre, We are proxying request from apache to tomcat using mod_proxy and mod_proxy_ajp Following is the load balancer configuration in apache. kindly let me know in case you need further details. # Port 80 Listen server_name:80 #Listen 153.88.86.250:80 VirtualHost _default_:80 ServerName server_name Timeout 10800 ProxyTimeout 10800 ProxyRequests Off ProxyPass /eMatrix balancer://cluster stickysession=JSESSIONID|jsessionid nofailover=On #ProxyPass /eMatrix balancer://cluster stickysession=JSESSIONID nofailover=On ProxyPreserveHost On ProxyPass /eMatrix ajp://server_name:8009/eMatrix ProxyPass /eMatrix ajp://server_name:8010/eMatrix ProxyPass /eMatrix ajp://server_name:8011/eMatrix ProxyPass /eMatrix ajp://server_name:9009/eMatrix ProxyPass /eMatrix ajp://server_name:9010/eMatrix ProxyPass /eMatrix ajp://server_name:9011/eMatrix ProxyPassReverse /eMatrix ajp://server_name:8009/eMatrix ProxyPassReverse /eMatrix ajp://server_name:8010/eMatrix ProxyPassReverse /eMatrix ajp://server_name:8011/eMatrix ProxyPassReverse /eMatrix ajp://server_name:9009/eMatrix ProxyPassReverse /eMatrix ajp://server_name:9010/eMatrix ProxyPassReverse /eMatrix ajp://server_name:9011/eMatrix Proxy balancer://cluster BalancerMember ajp://server_name:8009/eMatrix route=marsprod_rmiserver_1 loadfactor=33 retry=60 BalancerMember ajp://server_name:8010/eMatrix route=marsprod_rmiserver_2 loadfactor=33 retry=60 BalancerMember ajp://server_name:8011/eMatrix route=marsprod_rmiserver_3 loadfactor=33 retry=60 BalancerMember ajp://server_name:9009/eMatrix route=marsprod_rmiserver_4 loadfactor=33 retry=60 BalancerMember ajp://server_name:9010/eMatrix route=marsprod_rmiserver_5 loadfactor=33 retry=60 #load balancing performed based on number of user requests #ProxySet lbmethod=byrequests # Report server is for Report purpose only so balancing is required now. BalancerMember ajp://server_name:9011/eMatrix route=marsprod_rmiserver_6 loadfactor=33 retry=60 ProxySet lbmethod=byrequests /Proxy #Status page for balancer Location /balancer SetHandler balancer-manager Order Deny,Allow Deny from all Allow from all /Location ProxyStatus On Location /status SetHandler server-status Order Deny,Allow Deny from all Allow from all /Location ErrorLog /opt/web/apache/app/mxora/logs/error.log /VirtualHost # END port 80 --- On Tue, 11/16/10, André Warnier a...@ice-sa.com wrote: From: André Warnier
Buffer Overflow
I'm running Tomcat 6.0.29 I'm seeing this in my catalina.out log file. I've been able to narrow the problem to a single JSP page, and the problem only happens with a single user when that user is using Windows XP/FireFox 3.6.12. I've tested the page with XP/Firefox 3.6.3 and can't reproduce the issue. The JSP page reads and writes custom cookies. Any ideas on what might be happening, or on how to get more information (like what is actually in the header)? Nov 15, 2010 11:48:00 PM org.apache.jk.common.MsgAjp cpBytes SEVERE: Buffer overflow: buffer.len=8192 pos=8190 data=103 Nov 15, 2010 11:48:00 PM org.apache.jk.common.MsgAjp cpBytes SEVERE: Overflow java.lang.Throwable at org.apache.jk.common.MsgAjp.cpBytes(MsgAjp.java:196) at org.apache.jk.common.MsgAjp.appendByteChunk(MsgAjp.java:170) at org.apache.jk.common.MsgAjp.appendBytes(MsgAjp.java:155) at org.apache.jk.common.JkInputStream.appendHead(JkInputStream.java:324) at org.apache.jk.core.MsgContext.action(MsgContext.java:266) at org.apache.coyote.Response.action(Response.java:183) at org.apache.coyote.Response.sendHeaders(Response.java:379) at org.apache.catalina.connector.OutputBuffer.doFlush(OutputBuffer.java:305) at org.apache.catalina.connector.OutputBuffer.close(OutputBuffer.java:273) at org.apache.catalina.connector.Response.finishResponse(Response.java:493) at org.apache.catalina.connector.CoyoteAdapter.service(CoyoteAdapter.java:322) at org.apache.jk.server.JkCoyoteHandler.invoke(JkCoyoteHandler.java:190) at org.apache.jk.common.HandlerRequest.invoke(HandlerRequest.java:291) at org.apache.jk.common.ChannelSocket.invoke(ChannelSocket.java:774) at org.apache.jk.common.ChannelSocket.processConnection(ChannelSocket.java:703) at org.apache.jk.common.ChannelSocket$SocketConnection.runIt(ChannelSocket.java:896) at org.apache.tomcat.util.threads.ThreadPool$ControlRunnable.run(ThreadPool.java:690) at java.lang.Thread.run(Thread.java:619) Nov 15, 2010 11:48:00 PM org.apache.catalina.connector.CoyoteAdapter service SEVERE: An exception or error occurred in the container during the request processing java.lang.ArrayIndexOutOfBoundsException Nov 15, 2010 11:48:00 PM org.apache.jk.common.MsgAjp cpBytes SEVERE: Buffer overflow: buffer.len=8192 pos=8190 data=103 Nov 15, 2010 11:48:00 PM org.apache.jk.common.MsgAjp cpBytes SEVERE: Overflow java.lang.Throwable at org.apache.jk.common.MsgAjp.cpBytes(MsgAjp.java:196) at org.apache.jk.common.MsgAjp.appendByteChunk(MsgAjp.java:170) at org.apache.jk.common.MsgAjp.appendBytes(MsgAjp.java:155) at org.apache.jk.common.JkInputStream.appendHead(JkInputStream.java:324) at org.apache.jk.core.MsgContext.action(MsgContext.java:266) at org.apache.jk.core.MsgContext.action(MsgContext.java:302) at org.apache.coyote.Response.action(Response.java:183) at org.apache.coyote.Response.finish(Response.java:305) at org.apache.jk.server.JkCoyoteHandler.invoke(JkCoyoteHandler.java:195) at org.apache.jk.common.HandlerRequest.invoke(HandlerRequest.java:291) at org.apache.jk.common.ChannelSocket.invoke(ChannelSocket.java:774) at org.apache.jk.common.ChannelSocket.processConnection(ChannelSocket.java:703) at org.apache.jk.common.ChannelSocket$SocketConnection.runIt(ChannelSocket.java:896) at org.apache.tomcat.util.threads.ThreadPool$ControlRunnable.run(ThreadPool.java:690) at java.lang.Thread.run(Thread.java:619) Nov 15, 2010 11:48:00 PM org.apache.jk.common.ChannelSocket processConnection SEVERE: Error, processing connection java.lang.ArrayIndexOutOfBoundsException - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: the tomcat encoding
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Rob,
On 11/16/2010 4:31 AM, Rob Gregory wrote:
I also use a filter to set the following:-
request.setCharacterEncoding(UTF-8);
response.setContentType(text/html; charset=UTF-8);
response.setCharacterEncoding(UTF-8);
FWIW, you shouldn't override any character encoding sent by the client.
If there is no encoding, feel free to default to UTF-8.
Within the final html output I always include the following meta tag:-
meta http-equiv=Content-Type content=text/xhtml;
charset=utf-8/meta
This only makes sense if the response is actually in UTF-8. I would
recommend something like this:
meta http-equiv=Content-Type content=text/xhtml;
charset={response.characterEncoding} /
http://wiki.apache.org/tomcat/FAQ/CharacterEncoding
- -chris
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.10 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAkzi2RQACgkQ9CaO5/Lv0PBlaQCeOMxcnhPuX4hV/TPHT94+tmaM
oyMAoKR87o0U7voSaQ0OiUg9useP/aGJ
=O2Sy
-END PGP SIGNATURE-
-
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
Re: After manager says that there was a leak, how to use a profiler?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Brian, On 11/16/2010 10:42 AM, Brian wrote: If I choose the report of Paths from GC roots from it, I see LOTS of items! And all of them have the loader of indication. All the items are classes programmed by me, so I guess I'm guilty, and not the people that programmed Tomcat ;-) . They are DTOs, facades, Struts actionforms, etc. Why are all these classes in the list? If these are java.lang.Class objects, then you are looking at the list of classes loaded by the WebappClassLoader. These are the objects that make up the memory block that is unrecoverable. It's not the Class objects themselves that are holding onto references to the ClassLoader, it's the other way around. What should I check about them? Why are they still loaded? The real problem is that there is an object that is still live somewhere whose Class has been loaded by your webapp's WebappClassLoader. I guess there is something missing in my programming. Maybe I should do something about all the objects that are still alive that correspond to all the loaded classes? Maybe I need to do something in the contextDestroyed() method in the ServletContextListener in my app, in order to destroy all the objects properly? A basic rule is that anything you configure in a ServletContextListener's contextStarted method needs to be destroyed in the contextDestroyed method. Same thing for any Servlets that have init() methods -- use the destroy() method to clean-up those resources. When you undeploy, there may be a leak message in catalina.out. If you post that, we might be able to help you. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi3KwACgkQ9CaO5/Lv0PC/AwCgg9xb+4vGGrhp1DCVVUDrj5fL 03sAoJ4WjGXllEuymrPPf3BpAXPyMGzI =VNNp -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat Going down Frequently
On 16/11/2010 14:59, Amol Puglia wrote: Hello Andre, We are proxying request from apache to tomcat using mod_proxy and mod_proxy_ajp Following is the load balancer configuration in apache. kindly let me know in case you need further details. # Port 80 Listen server_name:80 #Listen 153.88.86.250:80 VirtualHost _default_:80 ServerName server_name Timeout 10800 ProxyTimeout 10800 ProxyRequests Off The next line: ProxyPass /eMatrix balancer://cluster stickysession=JSESSIONID|jsessionid nofailover=On #ProxyPass /eMatrix balancer://cluster stickysession=JSESSIONID nofailover=On ProxyPreserveHost On Overrides all of these lines, so they're no use: ProxyPass /eMatrix ajp://server_name:8009/eMatrix ProxyPass /eMatrix ajp://server_name:8010/eMatrix ProxyPass /eMatrix ajp://server_name:8011/eMatrix ProxyPass /eMatrix ajp://server_name:9009/eMatrix ProxyPass /eMatrix ajp://server_name:9010/eMatrix ProxyPass /eMatrix ajp://server_name:9011/eMatrix I don't think* this applies, if the above is wrong: ProxyPassReverse /eMatrix ajp://server_name:8009/eMatrix ProxyPassReverse /eMatrix ajp://server_name:8010/eMatrix ProxyPassReverse /eMatrix ajp://server_name:8011/eMatrix ProxyPassReverse /eMatrix ajp://server_name:9009/eMatrix ProxyPassReverse /eMatrix ajp://server_name:9010/eMatrix ProxyPassReverse /eMatrix ajp://server_name:9011/eMatrix p * Someone will correct me if I'm wrong Proxy balancer://cluster BalancerMember ajp://server_name:8009/eMatrix route=marsprod_rmiserver_1 loadfactor=33 retry=60 BalancerMember ajp://server_name:8010/eMatrix route=marsprod_rmiserver_2 loadfactor=33 retry=60 BalancerMember ajp://server_name:8011/eMatrix route=marsprod_rmiserver_3 loadfactor=33 retry=60 BalancerMember ajp://server_name:9009/eMatrix route=marsprod_rmiserver_4 loadfactor=33 retry=60 BalancerMember ajp://server_name:9010/eMatrix route=marsprod_rmiserver_5 loadfactor=33 retry=60 #load balancing performed based on number of user requests #ProxySet lbmethod=byrequests # Report server is for Report purpose only so balancing is required now. BalancerMember ajp://server_name:9011/eMatrix route=marsprod_rmiserver_6 loadfactor=33 retry=60 ProxySet lbmethod=byrequests /Proxy #Status page for balancer Location /balancer SetHandler balancer-manager Order Deny,Allow Deny from all Allow from all /Location ProxyStatus On Location /status SetHandler server-status Order Deny,Allow Deny from all Allow from all /Location ErrorLog /opt/web/apache/app/mxora/logs/error.log /VirtualHost # END port 80 --- On Tue, 11/16/10, André Warnier a...@ice-sa.com wrote: From: André Warnier a...@ice-sa.com Subject: Re: Tomcat Going down Frequently To: Tomcat Users List users@tomcat.apache.org Date: Tuesday, November 16, 2010, 4:55 PM Amol Puglia wrote: Hello Andre, Thanks for the updates. I am not using mod_jk , I am using proxy modules for redirecting request from apache to tomcat. Ok, sorry, I confused your post with another one. But anyway, your configuration shows a whole lot of proxy modules, many of which do not seem to have any relation with proxying calls to Tomcat. Such as these : LoadModule proxy_connect_module modules/mod_proxy_connect.so LoadModule proxy_ftp_module modules/mod_proxy_ftp.so LoadModule proxy_scgi_module modules/mod_proxy_scgi.so The following /may/ have a relation with the proxying to Tomcat, but which one are you really using ? LoadModule proxy_http_module modules/mod_proxy_http.so LoadModule proxy_ajp_module modules/mod_proxy_ajp.so Usually, people use *either* mod_proxy_http *or* mod_proxy_ajp, so which one is being used ? Can you provide some information about the apache httpd configuration that shows what is being proxied and how ? As you said that there is problem in application code, Is there anything we can point out where exactly is the problem in the code or configuration? I thought that Pid already pointed that out. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org 0x62590808.asc Description: application/pgp-keys signature.asc Description: OpenPGP digital signature
Re: Tomcat 6.0.29 Cookie.getMaxAge() always returns -1 regardless of actual value of maximum age
On 16/11/2010 16:58, Igor Barkon wrote: I run Tomcat 6.0.29 on Windows XP and IE 8.0.6001.18702. I noticed that getMaxAge() method of any cookie always returns -1 even for persistent cookie with real positive value of maximum age. IE treats this cookie correctly: cookie persists for amount of seconds specified by setMaxAge method. But getMaxAge always return -1. Based on specification getMaxAge must return actual value of maximum age of a given cookie, not default value. Is this a bug? How are you coming to this conclusion? What code are you using? What is setting the cookie? What happens in other browsers? What are the values of the relevant headers (Set-Cookie:, Cookie:) if you use a traffic inspection tool (Fiddler, Firebug, Chrome/Safari Developer)? p 0x62590808.asc Description: application/pgp-keys signature.asc Description: OpenPGP digital signature
Re: Tomcat 6.0.29 Cookie.getMaxAge() always returns -1 regardless of actual value of maximum age
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Igor, On 11/16/2010 11:58 AM, Igor Barkon wrote: I run Tomcat 6.0.29 on Windows XP and IE 8.0.6001.18702. I noticed that getMaxAge() method of any cookie always returns -1 even for persistent cookie with real positive value of maximum age. IE treats this cookie correctly: cookie persists for amount of seconds specified by setMaxAge method. But getMaxAge always return -1. Based on specification getMaxAge must return actual value of maximum age of a given cookie, not default value. Is this a bug? Are you checking the cookie you are generating on the server-side, or are you inspecting the cookie sent by the client afterward? Clients often do not send 100% of the cookie information back to the server with the request. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi4jgACgkQ9CaO5/Lv0PBpNwCfY1sifVzpczJODgPtBVPOFHj7 vXYAn07z64NXN1UcAUpfAiii2nfLu9ba =j3LO -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
RE: Using mod_jk in cluster environment responds HTTP 500
worker.list=loadbal_hub_prod,jkstatus,hub0,hub1 This is wrong. If the workers are already balanced (by the loadbal_hub_prod balancer), then they should not be separately listed in worker.list. Correct : worker.list=loadbal_hub_prod,jkstatus Well, assuming he doesn't have any mounts pointing specifically to hub0 or hub1. In that case, he's fine. -Tony - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
RE: Tomcat 6.0.29 using more and more RAM until it collapses?
Hu Pid, Now that I looked, I did find something! Nov 15, 2010 8:37:28 PM org.apache.catalina.loader.WebappClassLoader clearReferencesThreads SEVERE: The web application [] is still processing a request that has yet to finish. This is very likely to create a memory leak. You can control the time allowed for requests to finish by using the unloadDelay attribute of the standard Context implementation. Nov 15, 2010 8:37:28 PM org.apache.catalina.loader.WebappClassLoader clearReferencesThreads SEVERE: The web application [] is still processing a request that has yet to finish. This is very likely to create a memory leak. You can control the time allowed for requests to finish by using the unloadDelay attribute of the standard Context implementation. Nov 15, 2010 8:37:28 PM org.apache.catalina.loader.WebappClassLoader clearReferencesThreads SEVERE: The web application [] is still processing a request that has yet to finish. This is very likely to create a memory leak. You can control the time allowed for requests to finish by using the unloadDelay attribute of the standard Context implementation. Nov 15, 2010 8:37:29 PM org.apache.catalina.loader.WebappClassLoader loadClass INFO: Illegal access: this web application instance has been stopped already. Could not load com.manuals.common.utility.Log. The eventual following stack trace is caused by an error thrown for debugging purposes as well as to attempt to terminate the thread which caused the illegal access, and has no functional impact. java.lang.IllegalStateException at org.apache.catalina.loader.WebappClassLoader.loadClass(WebappClassLoader.java:1531) at org.apache.catalina.loader.WebappClassLoader.loadClass(WebappClassLoader.java:1491) at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:320) at com.manuals.model.utility.DatabaseObject.getStatement(DatabaseObject.java:434) at com.manuals.model.utility.DatabaseObject.getQueryResultSet(DatabaseObject.java:366) at com.manuals.model.utility.DatabaseObject.getQueryResultSet(DatabaseObject.java:326) And also this: ov 15, 2010 8:31:09 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:10 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:11 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:12 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:13 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:14 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:16 PM org.apache.catalina.loader.WebappClassLoader clearReferencesThreads SEVERE: The web application [] appears to have started a thread named [MySQL Statement Cancellation Timer] but has failed to stop it. This is very likely to create a memory leak. Nov 15, 2010 8:31:20 PM org.apache.catalina.core.StandardContext stop INFO: Container org.apache.catalina.core.ContainerBase.[Catalina].[localhost].[/] has not been started Nov 15, 2010 8:31:21 PM org.apache.catalina.startup.HostConfig checkResources INFO: Undeploying context [] -Original Message- From: Pid [mailto:p...@pidster.com] Sent: Tuesday, November 16, 2010 03:43 AM To: Tomcat Users List Subject: Re: Tomcat 6.0.29 using more and more RAM until it collapses? On 16/11/2010 05:08, Brian wrote: I still have to find the reason of the leak. Tomcat attempts to find notify you of potential memory leaks. You reported messages in your logs at app reload. Did you tell us what they are yet? p - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat 6.0.29 Cookie.getMaxAge() always returns -1 regardless of actual value of maximum age
Sorry, may be I use too many words to explain the problem. I just tried to be
more specific.
I run development. Everything (including Tomcat itself) is located in the same
computer. Tomcat is configured to port .
I used example software provided by Tomcat and located in Tomcat webapps
subdirectory. I just slightly modified it to scrutiny a little bit more on
cookies. I added max age as input for a new cookie (sent from Tomcat to IE),
and
I displayed this field in a list of cookies (sent to Tomcat from IE).
More specificaly: I created a new cookie with name, value and custom maximum
age, and sent it to Tomcat as a parameter; I sent this cookie to IE as a
cookie;
next time when I created a new cookie the old one was available for me as a
cookie and I displayed it on a screen to verify my entry.
I noticed that max age was always displayed as -1 even though it was originally
set to 60 and was treated by IE as 60 (cookie was disabled by IE in 60 seconds
after it was created, as it is supposed to be).
Fiddler was usless, it did not see my transactions (probably because Tomcat and
IE are located on the same computer). I used IE HTTP Analyzer v5.
Sincerely,
Igor
This is my servlet:
import javax.servlet.http.HttpServlet;
import java.io.IOException;
import javax.servlet.ServletException;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
import javax.servlet.RequestDispatcher;
import java.io.PrintWriter;
import java.util.ResourceBundle;
import javax.servlet.http.Cookie;
import javax.servlet.http.HttpSession;
public class CookieExampleIgor extends HttpServlet
{
ResourceBundle rb = ResourceBundle.getBundle(LocalStrings);
public void doGet(
HttpServletRequest request,
HttpServletResponse response
) throws IOException, ServletException
{
String cookieName;
String cookieValue;
int iMaxAge = -1;
String sMaxAge;
response.setContentType(text/html);
PrintWriter out = response.getWriter();
out.println(html);
out.println(head);
String title = rb.getString(cookies.title);
out.println(title + title + /title);
out.println(/head);
out.println(body);
out.println(form action=\CookieExampleIgor\ method=\post\);
out.println(a href=\../cookies.html\);
out.println(img src=\../images/code.gif\ height=24 +
width=24 align=right border=0 alt=\view code\/a);
out.println(a href=\../index.html\);
out.println(img src=\../images/return.gif\ height=24 +
width=24 align=right border=0 alt=\return\/a);
out.println(h3 + title + /h3);
out.println(This is Igor\'s job while investigating cookiesbr);
HttpSession hs = request.getSession(false);
if ((request.getParameter(sessionreset) != null) (hs != null))
{
hs.invalidate();
out.println(Session has been invalidatedbr);
};
if ((request.getParameter(sessioncreate) != null) (hs == null))
{
hs = request.getSession();
};
if(hs != null)
{
out.println(Session ID: + hs.getId() + br);
out.println(Session Max Inactive Interval: + hs.getMaxInactiveInterval()
+
br);
}
else
{
out.println(Session ID: nullbr);
}
cookieName = request.getParameter(cookiename);
cookieValue = request.getParameter(cookievalue);
sMaxAge = request.getParameter(maxage);
if (request.getParameter(submitcookie) != null cookieName != null
cookieValue != null
cookieName != cookieValue !=
)
{
if (sMaxAge != null)
{
try
{
iMaxAge = Integer.decode(sMaxAge);
}
catch (NumberFormatException e)
{
sMaxAge += ; NumberFormatException; Default value is used + iMaxAge;;
}
}
else
{
sMaxAge = Null value specified; Default value is used + iMaxAge;
}
Cookie newCookie = new Cookie(cookieName, cookieValue);
newCookie.setMaxAge(iMaxAge); // I set it to 60, and it was 60, IE
disabled it in 60 secs
response.addCookie(newCookie);
out.println(p);
out.println(rb.getString(cookies.set) + /br);
out.println(Name: + cookieName + /br);
out.println(Value: + cookieValue + br);
out.println(Max Age: + sMaxAge);
out.println(/p);
}
Cookie[] cookies = request.getCookies();
if (cookies != null cookies.length 0)
{
out.println(br + rb.getString(cookies.cookies));
for(int i = 0; i cookies.length; i++)
{
out.println(brCookie name: + cookies[i].getName());
out.println(nbsp;Cookie value: + cookies[i].getValue());
out.println(nbsp;Max Age: + cookies[i].getMaxAge()); // It is
always
-1 on display
out.println(nbsp;Path: + cookies[i].getPath());
}
}
else
{
out.println(br + rb.getString(cookies.no-cookies));
}
out.println(p);
out.println(rb.getString(cookies.make-cookie) + br);
out.println(Name:input type = \text\ name = \cookiename\br);
out.println(Value:input type = \text\ name = \cookievalue\br);
Re: Tomcat 6.0.29 Cookie.getMaxAge() always returns -1 regardless of actual value of maximum age
I created cookie as a parameter on IE and sent it to Tomcat. On server the parameter was transformed to a cookie and sent back to IE. Next time when IE sent new request to the server, the old cookie is part of Request (now it is cookie, nort parameter), so I can display this information to verify. I also believe that IE just sent incomplete information to Server, only name and value. From: Christopher Schultz ch...@christopherschultz.net To: Tomcat Users List users@tomcat.apache.org Sent: Tue, November 16, 2010 1:57:44 PM Subject: Re: Tomcat 6.0.29 Cookie.getMaxAge() always returns -1 regardless of actual value of maximum age -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Igor, On 11/16/2010 11:58 AM, Igor Barkon wrote: I run Tomcat 6.0.29 on Windows XP and IE 8.0.6001.18702. I noticed that getMaxAge() method of any cookie always returns -1 even for persistent cookie with real positive value of maximum age. IE treats this cookie correctly: cookie persists for amount of seconds specified by setMaxAge method. But getMaxAge always return -1. Based on specification getMaxAge must return actual value of maximum age of a given cookie, not default value. Is this a bug? Are you checking the cookie you are generating on the server-side, or are you inspecting the cookie sent by the client afterward? Clients often do not send 100% of the cookie information back to the server with the request. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi4jgACgkQ9CaO5/Lv0PBpNwCfY1sifVzpczJODgPtBVPOFHj7 vXYAn07z64NXN1UcAUpfAiii2nfLu9ba =j3LO -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat Going down Frequently
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 André, On 11/16/2010 1:21 PM, André Warnier wrote: WARNING: Exception processing manager org.apache.catalina.session.standardmana...@a20981 background process java.lang.IllegalStateException: getAttributeNames: Session already invalidated at org.apache.catalina.session.StandardSession.getAttributeNames(StandardSession.java:1052) at org.apache.catalina.session.StandardSessionFacade.getAttributeNames(StandardSessionFacade.java:120) at com.ericsson.mars.jspbean.LoginBean.logout(LoginBean.java:2366) at com.ericsson.mars.jspbean.LoginBean.valueUnbound(LoginBean.java:2450) Just to add a bit more for reference: at org.apache.catalina.session.StandardSession.removeAttributeInternal(StandardSession.java:1654) at org.apache.catalina.session.StandardSession.expire(StandardSession.java:756) at org.apache.catalina.session.StandardSession.isValid(StandardSession.java:592) at org.apache.catalina.session.ManagerBase.processExpires(ManagerBase.java:680) It looks like LoginBean implements HttpSessionBindingListener and it's being triggered during the expiration of the session. I actually don't see any reason why this code should cause an exception: 1. The valueUnbound method must be called after the value is inaccessible via HttpSession.getAttribute 2. There is no way for a client to check the validity of an HttpSession object before calling getAttribute (or setAttribute for that matter) The HttpSession interface javadoc says For session that are invalidated or expire, notifications are sent after the session has been invalidated or expired. which means that, at least, the state above is correct. The above is a WARNING message, indicating that Tomcat has caught something unusual. The important point there is that it is a WARNING. So, 3 things: 1. This is an exception, not some fatal condition 2. This exception is being caught a logged, so you shouldn't worry too much 3. This is logged as a WARNING, so you shouldn't worry too much Now, if this exception is actually causing the container to shut itself down, then it's a serious problem. Here's how you can check: 1. Start Tomcat normally 2. Log into your webapp 3. Log out of your webapp (this should trigger LoginBean.valueUnbound) 4. See if Tomcat shuts down If Tomcat stays running, then you're okay so far. Now: 1. Change the session timeout from whatever it is (default=30 minutes) to 1 minute 2. Start Tomcat normally 3. Log into your webapp 4. Wait 2 minutes (use a stopwatch) 5. Try to do something with your webapp If Tomcat stays running, then this exception is just a warning and your stability problems lie elsewhere. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi9HQACgkQ9CaO5/Lv0PAKUQCgtKUMQlT0Zy75blGbOi5K8Hce lEMAnR43PzXl79cADDCXZBgejk1e1cT7 =CCSL -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
RE: After manager says that there was a leak, how to use a profiler?
All my JARs are in the shared directory that Tomcat has. I don't upload them in the WAR everytime I deploye again my app. Is that a problem? I have just read somewhere do it could actually be a problem, that when they are shared among the apps, bad things happen. I have basically two apps running, but are EXACTLY the same WAR, but they run in different domain names. So they use the same JARs, and do exactly the same. The other 20 WARs are just full of static HTML pages. -Original Message- From: Mark Thomas [mailto:ma...@apache.org] Sent: Tuesday, November 16, 2010 04:37 AM To: Tomcat Users List Subject: Re: After manager says that there was a leak, how to use a profiler? On 16/11/2010 04:12, Brian wrote: Maybe my question sounded too vague and repetitive. What I meant is something like this Is the a class loader (or something like that) than I should start analyzing? I mean, is there a route I should follow in order to see which objects are stucked? If the Tomcat Manager knows that something got stucked in the memory, how does it know that? Is there a clue we should start?. http://markmail.org/message/fcbvwapt6afyndxn Mark - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
RE: After manager says that there was a leak, how to use a profiler?
From: Brian [mailto:bbprefix-m...@yahoo.com] Subject: RE: After manager says that there was a leak, how to use a profiler? All my JARs are in the shared directory that Tomcat has. Is that a problem? Yes, that's a problem. You very likely have some references in the shared classes that point to objects of one webapp or the other - and that will prevent their every being garbage collected. Put the jars where they belong - in each webapp's WEB-INF/lib directory. Disk is cheap, instability isn't. - Chuck THIS COMMUNICATION MAY CONTAIN CONFIDENTIAL AND/OR OTHERWISE PROPRIETARY MATERIAL and is thus for use only by the intended recipient. If you received this in error, please contact the sender and delete the e-mail and its attachments from all computers. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Using mod_jk in cluster environment responds HTTP 500
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Rikslovein, On 11/15/2010 12:47 AM, rikslovein wrote: Server.xml Connector port=8009 address=localhost enableLookups=false redirectPort=8443 protocol=AJP/1.3 / Is redirectPort=8443 correct for your environment? This has nothing to do with your HTTP 500 problems, but it's worth having a clean config. workers.properties: workers.tomcat_home=/opt/webhost/DWH-Tomcat-Inst/tomcat5.5.23 workers.java_home=/usr/lib/jvm/java-1.5.0 ps=/ Remove the previous 3 properties: they have no effect and clutter your configuration. worker.list=loadbal_hub_prod,jkstatus,hub0,hub1 I don't believe you have to mention hub0 and hub1 directly, here, but it doesn't hurt. worker.hub0.port=8009 worker.hub0.host=localhost worker.hub0.type=ajp13 worker.hub0.lbfactor=1 worker.hub0.redirect=hub1 worker.hub1.port=9009 worker.hub1.host=localhost worker.hub1.type=ajp13 worker.hub1.lbfactor=1 worker.hub1.redirect=hub0 Check out the worker.XXX.reference property: it will save you some typing and help maintain consistency across lb workers. What version of mod_jk are you currently using? Actually... what versions of /everything/ are you using? worker.jkstatus.type=status worker.loadbal_hub_prod.type=lb worker.loadbal_hub_prod.balanced_workers=hub0,hub1 worker.loadbal_hub_prod.sticky_session=false I notice that you don't have any strategy for determining availability of workers configured (ping_mode). You might want to look at those options to determine what is appropriate for your environment. I see that the sticky_session_force directive can result in a 500 error being returned to a user who is trying to access a balanced worker in an error state. I also see that you have sticky_session=false and the sticky_session_force setting should be irrelevant. Is it possible that your configuration file isn't actually being used? Try introducing a syntax error into workers.properties and seeing if mod_jk complains about it on startup. Can you show us what the status worker says when you find that 500 responses are being returned? - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi948ACgkQ9CaO5/Lv0PAzVACeJ86i2QDwkljprTrSEJ7Axlz9 63MAn2YBHgUfBpEqqi1Tz0dtk73A4n+r =ORgo -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Using mod_jk in cluster environment responds HTTP 500
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Rikslovein, On 11/15/2010 5:33 AM, rikslovein wrote: None of my tomcat is configured with HTTP connector. They are configured only for AJP. Does the jkmanager requires this HTTP port? Do you mean the manager webapp that comes with Tomcat? No, it doesn't require any HTTP connector. Any relation between the non-SSL HTTP connector and jkmanager? There is no relationship between any webapp and any connector that I know of. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi+AEACgkQ9CaO5/Lv0PAr1gCeI4fY5f0RyBSjGjxprzXlzGul /0sAn2EOuqT6qNR6lcD4YKNs0d0/mLCC =IyKY -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat 6.0.29 using more and more RAM until it collapses?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Brian, On 11/16/2010 3:23 PM, Brian wrote: Nov 15, 2010 8:37:28 PM org.apache.catalina.loader.WebappClassLoader clearReferencesThreads SEVERE: The web application [] is still processing a request that has yet to finish. This is very likely to create a memory leak. You can control the time allowed for requests to finish by using the unloadDelay attribute of the standard Context implementation. Looks like you have some (or at least one) long-running request. Maybe it's generating a ton of buffered output, too :) Nov 15, 2010 8:37:28 PM org.apache.catalina.loader.WebappClassLoader clearReferencesThreads SEVERE: The web application [] is still processing a request that has yet to finish. This is very likely to create a memory leak. You can control the time allowed for requests to finish by using the unloadDelay attribute of the standard Context implementation. Again. Nov 15, 2010 8:37:28 PM org.apache.catalina.loader.WebappClassLoader clearReferencesThreads SEVERE: The web application [] is still processing a request that has yet to finish. This is very likely to create a memory leak. You can control the time allowed for requests to finish by using the unloadDelay attribute of the standard Context implementation. Interesting that it generated all of these messages at the same time. I'm not sure if that means you have 3 requests pending or that 3 different problems were found with one thread. Nov 15, 2010 8:37:29 PM org.apache.catalina.loader.WebappClassLoader loadClass INFO: Illegal access: this web application instance has been stopped already. Could not load com.manuals.common.utility.Log. The eventual following stack trace is caused by an error thrown for debugging purposes as well as to attempt to terminate the thread which caused the illegal access, and has no functional impact. java.lang.IllegalStateException at org.apache.catalina.loader.WebappClassLoader.loadClass(WebappClassLoader.java:1531) at org.apache.catalina.loader.WebappClassLoader.loadClass(WebappClassLoader.java:1491) at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:320) at com.manuals.model.utility.DatabaseObject.getStatement(DatabaseObject.java:434) at com.manuals.model.utility.DatabaseObject.getQueryResultSet(DatabaseObject.java:366) at com.manuals.model.utility.DatabaseObject.getQueryResultSet(DatabaseObject.java:326) So that thread is actually doing something -- it hasn't just blocked indefinitely: your DatabaseObject class is causing the WebappClassLoader to attempt to load a new class that hadn't been previously loaded. Since WebappClassLoader knows it's shutting down (or has already shut down at that point), it is refusing to fulfill the request. And also this: ov 15, 2010 8:31:09 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:10 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:11 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:12 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:13 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Nov 15, 2010 8:31:14 PM org.apache.catalina.core.StandardWrapper unload INFO: Waiting for 1 instance(s) to be deallocated Since these messages are from earlier, I assume you were just pasting them in arbitrary order into your post: it looks like Tomcat waited for a while for your requests to complete (looping, complaining once per second) and then finally gave up and issued the messages above (...still processing a request...) at the end. Nov 15, 2010 8:31:16 PM org.apache.catalina.loader.WebappClassLoader clearReferencesThreads SEVERE: The web application [] appears to have started a thread named [MySQL Statement Cancellation Timer] but has failed to stop it. This is very likely to create a memory leak. There is almost nothing you can do about this except to upgrade your MySQL driver: http://bugs.mysql.com/bug.php?id=36565 Nov 15, 2010 8:31:20 PM org.apache.catalina.core.StandardContext stop INFO: Container org.apache.catalina.core.ContainerBase.[Catalina].[localhost].[/] has not been started This is a different issue. Note that [] (root) is not [/]. Have you specified your webapp's path somewhere. Don't do that. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi+qwACgkQ9CaO5/Lv0PAK6ACbBl2mBZnnwYFA5aw6GhA3bDxA kaEAoKhOgsCFMe6N0hgVrmLCMebU4OVw =sklc -END PGP SIGNATURE- - To unsubscribe, e-mail:
Re: Tomcat 6.0.29 Cookie.getMaxAge() always returns -1 regardless of actual value of maximum age
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Igor, On 11/16/2010 3:40 PM, Igor Barkon wrote: I created cookie as a parameter on IE and sent it to Tomcat. On server the parameter was transformed to a cookie and sent back to IE. Next time when IE sent new request to the server, the old cookie is part of Request (now it is cookie, nort parameter), so I can display this information to verify. I also believe that IE just sent incomplete information to Server, only name and value. I believe this is all that the HTTP session specification requires. See http://www.faqs.org/rfcs/rfc2109.html section 4.3.4 or http://www.faqs.org/rfcs/rfc2965.html section 3.3.4. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi/EUACgkQ9CaO5/Lv0PAGkgCdEvXAF1pDDozFzC6CFt01fgc0 oBkAnj1Jy+Fsgh+yLyyGeh6jK/w212S6 =AKLz -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Usage of Tomcat SSL in client-side class
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Moley, On 11/16/2010 9:33 AM, Moley Harey wrote: My question is what do I have to do in my Java client-side classes to work with HTTPS requests [to] Tomcat? Do I have to accept somehow the certificate provided by Tomcat in my Java classes? Yes. If you are using a certificate signed by a well-known CA (VeriSign, Thawte, etc.) then you probably don't need to do anything special. If you are using a self-signed certificate, you'll need to import that certificate into your certificate store (cacerts) and make sure that your client-side classes are configured to use it. Presumably, your web services client library makes this kind of thing easy. You'll, of course, have to use an https://; URL when trying to connect as well. Presumably, your web services client library makes this kind of thing easy. Well, I'm totally newbie in these security issues :-/ At least you're asking :) - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi/dkACgkQ9CaO5/Lv0PAMsgCfSV2S8E8leRxUlAMZnSnAuFRM wF8An2wUwuj4a6LnFZ/ZMB27saseiI2V =sbPz -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Buffer Overflow
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Tyson, First, let me say that I know almost nothing about the AJP protocol, but reading-through the code from the inside out has triggered a memory. See below. On 11/16/2010 1:26 PM, Tyson Lowery wrote: I'm running Tomcat 6.0.29 Thanks! I'm seeing this in my catalina.out log file. I've been able to narrow the problem to a single JSP page, and the problem only happens with a single user when that user is using Windows XP/FireFox 3.6.12. I've tested the page with XP/Firefox 3.6.3 and can't reproduce the issue. The JSP page reads and writes custom cookies. Can you give us some examples? Is it possible that you are generating very large cookie names and/or values? Any ideas on what might be happening, or on how to get more information (like what is actually in the header)? You ought to be able to configure the logger to emit messages for the org.apache.jk package tree. It looks like the code is generating debug messages if the logging system is configured to capture them. Nov 15, 2010 11:48:00 PM org.apache.jk.common.MsgAjp cpBytes SEVERE: Buffer overflow: buffer.len=8192 pos=8190 data=103 8192 is the default packetSize option for the AJP Connector. It appears that you are trying to shove more data into a single packet than is possible. It's not your fault. Well, sort-of, but not really. Your code is probably fine. I believe that either the AJP protocol or Tomcat's implementation of it requires that all of the HTTP headers fit into the first AJP packet sent back to the web server. So, if you have more than 8192 bytes of headers to send back to the web server, you'll get this error. That's what I asked about the large cookie values. It also happens when you use SSL with proxied certificate chains which can eat up a lot of space, too. Anyhow, there is good news: you can change the default packetSize like this: Connector ... packetSize=16384 / MAKE SURE YOU SET THE max_packet_size IN mod_jk TO BE THE SAME VALUE, otherwise everything will go to hell. You can go up to 65536 bytes for the packet size if you really need to. Note that larger packet sizes may have a performance impact. Another option would be to limit the total size of cookie information you allow your custom JSP to emit. It's not foolproof (you can always use some other header to reach this limit) but it might help if you are not comfortable increasing the AJP packet size for some reason. Hope that helps, - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzjAYMACgkQ9CaO5/Lv0PAsLACdHmU8yeG26CeiFn8I7pBVXB6R MBQAn1tUNC/aWasAKHuBhJ9EJAMQKonO =pjT3 -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat log files
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Tobias, On 11/15/2010 10:43 AM, Tobias Crefeld wrote: Am Tue, 09 Nov 2010 16:06:14 -0500 schrieb Christopher Schultz ch...@christopherschultz.net: Have you correctly replaced lib/tomcat-juli.jar and installed lib/tomcat-juli-adapters.jar? If not, you may be falling back to the old logger which requires logging.properties. For whatever-reason and different than tomcat-juli-adapters.jar the default location of tomcat-juli.jar is bin/, not lib/. Thanks for pointing that out. I had been reading the docs and didn't notice that they were different tomcat-juli.jar goes into bin/ while tomcat-juli-adapters.jar goes into lib/). - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzjAsIACgkQ9CaO5/Lv0PASzgCfVInqw5ASMbCoTGrLgFc2pcgS 5YMAn3S2YhSNRY/+9Rped7kixFOwdnnb =lPVN -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Tomcat Going down Frequently
On 16/11/2010 21:15, Christopher Schultz wrote: 2. There is no way for a client to check the validity of an HttpSession object before calling getAttribute (or setAttribute for that matter) The HttpSession interface javadoc says For session that are invalidated or expire, notifications are sent after the session has been invalidated or expired. which means that, at least, the state above is correct. The class is calling: StandardSessionFacade.getAttributeNames. Shouldn't we see the HttpSession class in the stacktrace, not the implementation, if that's what they're calling? Makes me think the OP is interacting with Tomcat internal classes. p 0x62590808.asc Description: application/pgp-keys signature.asc Description: OpenPGP digital signature
RE: After manager says that there was a leak, how to use a profiler?
Thanks a lot for the info Konstantin. I will read it asap! -Original Message- From: Konstantin Kolinko [mailto:knst.koli...@gmail.com] Sent: Monday, November 15, 2010 11:08 PM To: Tomcat Users List Subject: Re: After manager says that there was a leak, how to use a profiler? 2010/11/16 Brian bbprefix-m...@yahoo.com: Hi, After the Tomcat manager warns that there was a leak with one app what was stopped or undeployed, how do I use the profiler to investigate that? I mean, that is the starting point to analize that? What should I look for? E.g., using Eclipse MAT, [1] http://www.eclipse.org/mat/ [2] http://dev.eclipse.org/blogs/memoryanalyzer/2008/05/17/the-unknown- generation-perm/ Best regards, Konstantin Kolinko - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: the tomcat encoding
hi:
thanks, the problem is that my tomcat interact with other server and the
server use somekind of encoding that is different.
and after the encoding is transfered .it works now
thanks
2010/11/17 Christopher Schultz ch...@christopherschultz.net
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Rob,
On 11/16/2010 4:31 AM, Rob Gregory wrote:
I also use a filter to set the following:-
request.setCharacterEncoding(UTF-8);
response.setContentType(text/html; charset=UTF-8);
response.setCharacterEncoding(UTF-8);
FWIW, you shouldn't override any character encoding sent by the client.
If there is no encoding, feel free to default to UTF-8.
Within the final html output I always include the following meta tag:-
meta http-equiv=Content-Type content=text/xhtml;
charset=utf-8/meta
This only makes sense if the response is actually in UTF-8. I would
recommend something like this:
meta http-equiv=Content-Type content=text/xhtml;
charset={response.characterEncoding} /
http://wiki.apache.org/tomcat/FAQ/CharacterEncoding
- -chris
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.10 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAkzi2RQACgkQ9CaO5/Lv0PBlaQCeOMxcnhPuX4hV/TPHT94+tmaM
oyMAoKR87o0U7voSaQ0OiUg9useP/aGJ
=O2Sy
-END PGP SIGNATURE-
-
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org
Placing JARs in ..../tomcat/shared/lib causes leaks when stopping apps?
Hi, I'm using Tomcat 6.0.29. I have two identical apps running at the same time. I mean, they use exactly the same WAR, but each one runs in a different web domain. When I stop/redeploy one of the apps, very often I get the famous leak problem, because of the webclassloader not being able to get collected by the garbage collector. etc etc etc. I'm placing all my library JARs in a common directory (../tomcat/shared/lib), and I set the catalina.properties file so Catalina knows this directory exists. Why did I do that? For two reasons: 1- uploading a WAR without all the support JARs is faster than with them, and I'm uploading my WARs very often, because I'm changing my app constantly. 2- I THOUGHT/THINK that placing the JARs in the same shared directory is better, because it takes less memory, and because after all there must be a reason why the shared directory option existed by default (at least until Tomcat 6.0.14 came out). And given that my two apps use exactly the same WAR, it made sense to me to place the JARs in the shared dir. Now I have read that this issue may be the reason of my problem with leaks. Something like this for each app, the webapploader pretendes to unload all the classes, and if they are shared it is not possible and then the leak happens. Is that true? Should I avoid using the shared lib directory? If so, for what was this option created for? Thanks, Brian
RE: Placing JARs in ..../tomcat/shared/lib causes leaks when stopping apps?
From: Brian [mailto:bbprefix-m...@yahoo.com] Subject: Placing JARs in /tomcat/shared/lib causes leaks when stopping apps? Something like this for each app, the webapploader pretendes to unload all the classes, and if they are shared it is not possible and then the leak happens. Well, the wording of the above is rather odd; there's no pretending going on, and nothing can unload a class directly - that's done only by GC after the class becomes unused. Is that true? Sort of. Should I avoid using the shared lib directory? In almost all cases, yes. It's safe to use the shared library when the classes placed therein truly are library functions (no persistent objects, no keeping of references to their callers), or when an object *must* be shared across multiple webapps. If so, for what was this option created for? To allow /objects/ to be shared across webapps, in the rare instance that it's necessary. Webapps are intended to be independent of one another, and any mingling of their content limits that independence severely - as you are observing. Keep the webapps separate; it's a lot easier on everyone concerned. - Chuck THIS COMMUNICATION MAY CONTAIN CONFIDENTIAL AND/OR OTHERWISE PROPRIETARY MATERIAL and is thus for use only by the intended recipient. If you received this in error, please contact the sender and delete the e-mail and its attachments from all computers. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
RE: Placing JARs in ..../tomcat/shared/lib causes leaks when stopping apps?
Thanks chuck, now I understand it. My JARs are the regular type of JAR, that doesn't have to be shared by all the apps for some reason. I didn't understand the real pourpose of that directory indeed. I will change it now. -Original Message- From: Caldarale, Charles R [mailto:chuck.caldar...@unisys.com] Sent: Tuesday, November 16, 2010 11:06 PM To: Tomcat Users List Subject: RE: Placing JARs in /tomcat/shared/lib causes leaks when stopping apps? From: Brian [mailto:bbprefix-m...@yahoo.com] Subject: Placing JARs in /tomcat/shared/lib causes leaks when stopping apps? Something like this for each app, the webapploader pretendes to unload all the classes, and if they are shared it is not possible and then the leak happens. Well, the wording of the above is rather odd; there's no pretending going on, and nothing can unload a class directly - that's done only by GC after the class becomes unused. Is that true? Sort of. Should I avoid using the shared lib directory? In almost all cases, yes. It's safe to use the shared library when the classes placed therein truly are library functions (no persistent objects, no keeping of references to their callers), or when an object *must* be shared across multiple webapps. If so, for what was this option created for? To allow /objects/ to be shared across webapps, in the rare instance that it's necessary. Webapps are intended to be independent of one another, and any mingling of their content limits that independence severely - as you are observing. Keep the webapps separate; it's a lot easier on everyone concerned. - Chuck THIS COMMUNICATION MAY CONTAIN CONFIDENTIAL AND/OR OTHERWISE PROPRIETARY MATERIAL and is thus for use only by the intended recipient. If you received this in error, please contact the sender and delete the e-mail and its attachments from all computers. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Using mod_jk in cluster environment responds HTTP 500
Chris all, Thanks for your help and your responses once again. I now solved the problem. The problem was with the jkmanager URL I was using in the code. I think there is a change in URL attributes from older version. The current version uses vwa, vwr and vwf (which was coded as wa, wr wf respectively). I'm not sure on this but i saw them working on Unix where we have quite older version of Apache and Tomcat installed. Please confirm on this if you have any info. The jkmanager URL i was using previously and the changed URL is given below: Older URL (Works in Unix): - http://localhost/jkmanager/?cmd=updatew=loadbal_hub_prodsw=hub1wf=1wr=hub0wa=2mime=txt Changed URL (Working fine now on Linux): - http://localhost/jkmanager/?cmd=updatew=loadbal_hub_prodsw=hub1vwf=1vwr=hub0vwa=2mime=txt Thanks a lot again. Regards. Christopher Schultz-2 wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Rikslovein, On 11/15/2010 5:33 AM, rikslovein wrote: None of my tomcat is configured with HTTP connector. They are configured only for AJP. Does the jkmanager requires this HTTP port? Do you mean the manager webapp that comes with Tomcat? No, it doesn't require any HTTP connector. Any relation between the non-SSL HTTP connector and jkmanager? There is no relationship between any webapp and any connector that I know of. - -chris -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.10 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkzi+AEACgkQ9CaO5/Lv0PAr1gCeI4fY5f0RyBSjGjxprzXlzGul /0sAn2EOuqT6qNR6lcD4YKNs0d0/mLCC =IyKY -END PGP SIGNATURE- - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org -- View this message in context: http://old.nabble.com/Using-mod_jk-in-cluster-environment-responds-HTTP-500-tp30198809p30235421.html Sent from the Tomcat - User mailing list archive at Nabble.com. - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Request count increases drastically in tomcat 6.0.29
Hi, I am using an application in tomcat 6.0.29 and the java version is 1.6 64 bit. I am facing the problem in my application when the request count is more(this i am watching through Lambda probe). Please give me some idea why this request count is keeps increasing and how to destroy the idle requests.Herewith i attached the screenshot of that application --rujin - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
Re: Placing JARs in ..../tomcat/shared/lib causes leaks when stopping apps?
On Tue, 16 Nov 2010 22:55:10 -0500, Brian bbprefix-m...@yahoo.com wrote: I have two identical apps running at the same time. I mean, they use exactly the same WAR, but each one runs in a different web domain. [...] You should rather change your application to be aware of domain it runs in. Done that, there would be only one webapp deployed - no matter how many domains you need to handle. -- Mikolaj Rydzewski m...@ceti.pl - To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
