On 02/08/2024 12:06, Rainer Jung wrote:
Am 02.08.24 um 10:58 schrieb Mark Thomas:
On 02/08/2024 07:58, Mark Thomas wrote:
On 01/08/2024 23:48, Rainer Jung wrote:
<snip/>
I did not check each occurrence but here are examples which all also
have a NullPointer in the access log. I don't know whether that
triggers the failure or is just another symptom triggered by the
same root cause.
<snip/>
Rainer,
Thanks for the work you have put into this.
The combination of NPE and the timing of these emerging makes me
think the change to recycle the request and response may be the cause.
I'm holding off closing the 11.0.x vote while I look into this.
I can confirm the request/response recycling for HTTP/2 triggered
this. Reviewing the code with fresh eyes I see multiple issues. I'll
have fixes ready shortly but I am going to cancel the 11.0.0-M23
release vote. I am also going to change my vote to -1 for 10.1.27 and
9.0.92.
Thanks for investigating so quickly and thoroughly!
Unfortunately my testing zoo always needs a few days before test results
become reliable enough, so it is hard for me to report very quickly
after the start of the vote.
No problem. It is my fault for not spotting the problem when I was doing
the recycling changes.
The 11.0.x tests look good. I am going to back port now.
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
For additional commands, e-mail: dev-h...@tomcat.apache.org
