Re: OCSP stapling broken with 1.15.4

2018-10-01 Thread A. Schulze
Am 01.10.18 um 15:43 schrieb Bernardo Donadio: > I've restored the 1.15.4 package and have been making some requests. > Some of them are correctly stapled, others do not. There's no restart > between tests. maybe you run multiple threads and for each thread there is one first request? > I'm no

RE: OCSP stapling broken with 1.15.4

2018-10-01 Thread Reinis Rozitis
> Indeed, with further tests I think that the stapling is working... > sometimes. > > > I'm not using the staple file, though. Is this behavior expected without such > configuration? Also, I've enabled ssl_early_data. Each nginx worker has it's own cache. Depending on your worker_processes you m

Re: OCSP stapling broken with 1.15.4

2018-10-01 Thread Bernardo Donadio
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 10/1/18 10:04 AM, A. Schulze wrote: > Did you try to measure twice? Indeed, with further tests I think that the stapling is working... sometimes. I've restored the 1.15.4 package and have been making some requests. Some of them are correctly sta

Re: OCSP stapling broken with 1.15.4

2018-10-01 Thread A. Schulze
Bernardo Donadio: Hi. I've noticed that OCSP stapling was broken by 1.15.4, as you may see below: -- nginx 1.15.4 with OpenSSL 1.1.1 final $ openssl s_client -connect bcdonadio.com:443 -tlsextdebug -status CONNECTED(0003) TLS server extension "renegotiation info" (id=652

OCSP stapling broken with 1.15.4

2018-10-01 Thread Bernardo Donadio
Hi. I've noticed that OCSP stapling was broken by 1.15.4, as you may see below: -- nginx 1.15.4 with OpenSSL 1.1.1 final $ openssl s_client -connect bcdonadio.com:443 -tlsextdebug -status CONNECTED(0003) TLS server extension "renegotiation info" (id=65281), len=1 - 00