On 9/17/2018 2:20 PM, Rainer Jung wrote: > Am 17.09.2018 um 20:59 schrieb Daniel Ruggeri: >> Hi, all; >> I have been delayed executing the automation because the test >> suite seems to be hanging for me. This appears to be consistently >> during t/ssl/varlookup.t as that is the only process other than httpd >> running in this container during the hang. When killing httpd, the >> failures reported start at test 35 with err "Failed test 35 in >> t/ssl/varlookup.t at line 109 fail #35" and continue to 83. > > Which platform, which version of OpenSSL? >
Hi, Rainer; Sorry - I know it wasn't a very good report. I was just seeing if anyone has experienced a similar holdup. In fact, I let it run while tending to other things and came back to see it had completed (but failed), so perhaps it's not hung, but rather very slow. I'm using Debian 9 (stretch) in a container running on a 3.16.51-3 kernel. OpenSSL 1.1.1 is inside as well as several other dependencies... these are the "latests" that my build scripts grabbed: system: kernel: name: Linux release: 3.16.0-4-amd64 version: #1 SMP Debian 3.16.51-3 (2017-12-13) machine: x86_64 libraries: openssl: "1.1.1" openldap: "2.4.46" apr: "1.6.5" apr-util: "1.6.1" iconv: "1.2.2" brotli: "1.0.5" nghttp2: "1.33.0" zlib: "1.2.11" pcre: "8.42" libxml2: "2.9.8" php: "5.6.38" lua: "5.3.5" curl: "7.61.1" ...which reminds me... time to update my kernel :-) >> If anyone has experience with this area of the test suite, >> pointers definitely welcome. Otherwise, I'll start poking at it. > > You can run the tests with > > t/TEST -v -order=repeat > > which might give more insight. > > My passing tests for 2.4.34 e.g. logged > > ... > # testing : SSL_CIPHER_EXPORT > # expected: 'false' > # received: 'false' > ok 35 > # testing : SSL_CIPHER_ALGKEYSIZE > # expected: qr/^\d+$/ > # received: '128' > ok 36 > # testing : SSL_CIPHER_USEKEYSIZE > # expected: qr/^\d+$/ > # received: '128' > ok 37 > # testing : SSL_SECURE_RENEG > # expected: qr/^(false|true)$/ > # received: 'true' > ok 38 > ... > > and the comments always belong to the "ok" or "not ok" line that > follows the comments. So test 35 was the "SSL_CIPHER_EXPORT" test at > that time. I doubt it actually has to do with that. > > You might look at the thread stacks (with the gdb "bt" command or > similar) of the hanging httpd processes to gather more info. > > Any errors in the httpd error log, like "deadlock detected" etc.? Nothing stood out, but I did not debug too deeply. I'll let this next test run completely and do some poking around. Perhaps I was just impatient... but I don't recall the suite taking that long to run. Depending on how this goes, I may pause T&R until I can confirm an issue on my test rig, the server or the test code. If anyone else is able to build and verify that TLS w/ 2.4.x branch works A-OK for them, I'd be fine with proceeding - just being abundantly cautious. > > Regards, > > Rainer -- Daniel Ruggeri