shinrich commented on pull request #7134:
URL: https://github.com/apache/trafficserver/pull/7134#issuecomment-684934054
Looking at the code a bit more, I think this default takeover of the
inactivity timeout could occur for regular (non blind tunneled) connections as
well. The code in question that sets the inactivity timeout in
InactivityCop::check_inactivity was added somewhat recently by PR #6755. The
code in question
```
// set a default inactivity timeout if one is not set
if (ne->next_inactivity_timeout_at == 0 &&
nh.config.default_inactivity_timeout > 0) {
Debug("inactivity_cop", "vc: %p inactivity timeout not set, setting
a default of %d", ne,
nh.config.default_inactivity_timeout);
ne->set_default_inactivity_timeout(HRTIME_SECONDS(nh.config.default_inactivity_timeout));
NET_INCREMENT_DYN_STAT(default_inactivity_timeout_applied_stat);
}
```
Though if in fact next_inactivity_timeout_at and inactivity_timeout_in are
cleared by write_disable/read_disable, I don't see a way to get those reset
without the HttpSM level logic explicitly resetting it. And most of the
read/write_disable seems to occur at the NetVC level.
@sudheerv since you last touched the default_inactivity_timeout logic, do
you have any thoughts here?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
