Re: MaxInactiveInterval

2016-11-06 Thread Mark Thomas
On 06/11/2016 11:24, André Warnier (tomcat) wrote:
> On 04.11.2016 20:06, Mark Thomas wrote:



>> I did say "patches welcome" to André but since he is a committer that
>> should have been "commits welcome" ;)
>>
> 
> Guys, you all know my level of (in)competence in matters deeply Java.

You might be surprised at just how little Java I knew when I was made a
committer.

> I got my kudos by trying to help people in a general sense here, but
> when things seem to involve specific parts deep down the Tomcat code, I
> have to call for help.

As always, happy to provide pointers. This should end up as a docs fix
which I'd encourage you to consider tackling yourself.

> What I know is this, from the Servlet Specs 3.0 final :
> 
> quote
> 
> 12. session-config Element
> 
> The session-config defines the session parameters for this Web application.
> The sub-element *session-timeout* defines the default session time out
> interval
> for all sessions created in this Web application. The specified time out
> must be
> expressed in a whole number of minutes. If the time out is 0 or less,
> the container
> ensures the default behavior of sessions is never to time out. If this
> element is not
> specified, the container must set its default time out period.
> 
> unquote
> 
> So it appears that there is a difference between :
> - the WEB-INF/web.xml of a webapp specifying a session-timeout > 0, in
> minutes
> - the WEB-INF/web.xml of a webapp specifying a session-timeout =< 0
> (meaning, no timeout or "infinite")
> - the WEB-INF/web.xml of a webapp not specifying a session-timeout
> (container should supply a default value)
> 
> Which thus raises the question : if a web application does not set the
> session-timeout, what value is returned by Tomcat for
> getMaxInactiveInterval() ?
> 
> And the auxiliary question : can this (default) value be set somewhere
> in the configuration, or is this set in code ?
> 
> (Or is that the one that is set in (tomcat)/conf/web.xml :
> 
> 30
> 
> ?)

You are correct. The value from conf/web.xml provides the default if the
application's web.xml does not provide one.

> (and what happens if we remove that stanza from conf/web.xml ?)

There is a hard-coded default of 30 minutes (i.e. the same as the
explicit default in conf/web.xml) in o.a.catalina.core.StandardConetxt.

All of the above assumes that the standard components are being used.
Custom session managers, session implementations and/or Contexts could
all modify this behaviour.

To add to the fun, there is, effectively, an undocumented attribute on
Context - sessionTimeout - that could also be set in server.xml /
context.xml. Overall, the order of precedence should be:

- application specific web.xml
- conf/web.xml default
- sessionTimeout attribute on Context
- hard-coded default.

Mark


-
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org



Re: MaxInactiveInterval

2016-11-06 Thread tomcat

On 04.11.2016 20:06, Mark Thomas wrote:

On 04/11/2016 15:53, Caldarale, Charles R wrote:

From: Mark Thomas [mailto:ma...@apache.org] Subject: Re:
MaxInactiveInterval



On 04/11/2016 15:07, André Warnier (tomcat) wrote:

A log message in an application running under Tomcat 8 mentions
the "MaxInactiveInterval" setting, saying that it is a bit
short..

I think that I understand the meaning of the setting (the time
for which a session remains valid, even without interactions).
But where in the webapp context, and under what name, should an
equivalent parameter be specified ?



Set it in web.xml. Same with distributable.


This really appears to be a problem in the servlet spec (not
surprising), which mentions the getMaxInactiveInterval() and
setMaxInactiveInterval() APIs but only vaguely ties them to the
session-timeout element of session-config.  Tomcat doc does not
normally describe the nuances of the servlet spec, but something that
associates the config name with the API might be useful.


To be fair, Tomcat isn't helping here. We used to have these attributes
on the (session) manager but we removed with the expectation that they
would be configured in web.xml. We could have done a better job of
pointing to the expected locations.

I did say "patches welcome" to André but since he is a committer that
should have been "commits welcome" ;)



Guys, you all know my level of (in)competence in matters deeply Java.
I got my kudos by trying to help people in a general sense here, but when things seem to 
involve specific parts deep down the Tomcat code, I have to call for help.


What I know is this, from the Servlet Specs 3.0 final :

quote

12. session-config Element

The session-config defines the session parameters for this Web application.
The sub-element *session-timeout* defines the default session time out interval
for all sessions created in this Web application. The specified time out must be
expressed in a whole number of minutes. If the time out is 0 or less, the 
container
ensures the default behavior of sessions is never to time out. If this element 
is not
specified, the container must set its default time out period.

unquote

So it appears that there is a difference between :
- the WEB-INF/web.xml of a webapp specifying a session-timeout > 0, in minutes
- the WEB-INF/web.xml of a webapp specifying a session-timeout =< 0 (meaning, no timeout 
or "infinite")
- the WEB-INF/web.xml of a webapp not specifying a session-timeout (container should 
supply a default value)


Which thus raises the question : if a web application does not set the session-timeout, 
what value is returned by Tomcat for getMaxInactiveInterval() ?


And the auxiliary question : can this (default) value be set somewhere in the 
configuration, or is this set in code ?


(Or is that the one that is set in (tomcat)/conf/web.xml :

30

?)
(and what happens if we remove that stanza from conf/web.xml ?)





-
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org



Re: MaxInactiveInterval

2016-11-04 Thread Mark Thomas
On 04/11/2016 15:53, Caldarale, Charles R wrote:
>> From: Mark Thomas [mailto:ma...@apache.org] Subject: Re:
>> MaxInactiveInterval
> 
>> On 04/11/2016 15:07, André Warnier (tomcat) wrote:
>>> A log message in an application running under Tomcat 8 mentions
>>> the "MaxInactiveInterval" setting, saying that it is a bit
>>> short..
>>> 
>>> I think that I understand the meaning of the setting (the time
>>> for which a session remains valid, even without interactions). 
>>> But where in the webapp context, and under what name, should an 
>>> equivalent parameter be specified ?
> 
>> Set it in web.xml. Same with distributable.
> 
> This really appears to be a problem in the servlet spec (not
> surprising), which mentions the getMaxInactiveInterval() and
> setMaxInactiveInterval() APIs but only vaguely ties them to the
> session-timeout element of session-config.  Tomcat doc does not
> normally describe the nuances of the servlet spec, but something that
> associates the config name with the API might be useful.

To be fair, Tomcat isn't helping here. We used to have these attributes
on the (session) manager but we removed with the expectation that they
would be configured in web.xml. We could have done a better job of
pointing to the expected locations.

I did say "patches welcome" to André but since he is a committer that
should have been "commits welcome" ;)

Mark

-
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org



RE: MaxInactiveInterval

2016-11-04 Thread Caldarale, Charles R
> From: Mark Thomas [mailto:ma...@apache.org] 
> Subject: Re: MaxInactiveInterval

> On 04/11/2016 15:07, André Warnier (tomcat) wrote:
> > A log message in an application running under Tomcat 8 mentions the
> > "MaxInactiveInterval" setting, saying that it is a bit short..
> > 
> > I think that I understand the meaning of the setting (the time for which
> > a session remains valid, even without interactions).
> > But where in the webapp context, and under what name, should an
> > equivalent parameter be specified ?

> Set it in web.xml. Same with distributable.

This really appears to be a problem in the servlet spec (not surprising), which 
mentions the getMaxInactiveInterval() and setMaxInactiveInterval() APIs but 
only vaguely ties them to the session-timeout element of session-config.  
Tomcat doc does not normally describe the nuances of the servlet spec, but 
something that associates the config name with the API might be useful.

 - 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: MaxInactiveInterval

2016-11-04 Thread Mark Thomas
On 04/11/2016 15:07, André Warnier (tomcat) wrote:
> Hi.
> 
> A log message in an application running under Tomcat 8 mentions the
> "MaxInactiveInterval" setting, saying that it is a bit short..
> 
> The only place in the Tomcat 8 documentation where I find this setting,
> is in
> https://tomcat.apache.org/tomcat-8.0-doc/config/manager.html
> where it says that this is deprecated, and should be set in the Context.
> 
> However, the page at :
> https://tomcat.apache.org/tomcat-8.0-doc/config/context.html
> does not mention that parameter at all.
> 
> Deep puzzlement.
> What is this all about ?
> 
> I think that I understand the meaning of the setting (the time for which
> a session remains valid, even without interactions).
> But where in the webapp context, and under what name, should an
> equivalent parameter be specified ?

Set it in web.xml. Same with distributable.

Docs patches welcome.

Mark

-
To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
For additional commands, e-mail: users-h...@tomcat.apache.org