Taking this for a fix, it breaks every third travis run. 😠

Thanks @eiri, was going to get to it tomorrow if you didn't.

I’m having a stab at it at the moment.

@janl ok, I'll stop my work and switch to something different. Was self-assignment on this case not sufficient to indicate I'm already working on it?

@eiri sorry, I thought your clock ran out and the end of your day yesterday :/

@eiri I felt a bit over-eager because I had approved this PR and I should have caught the issue on review.

Closing this in favour of #1250. Apologies again for the toe-stepping.

@janl No worries, it's actually my bias of assuming me been online miracly obvious to everyone and I do seat in a weird timezone. Sorry this came across that grouchy.

All good :)

@janl Your fix failed on Jenkins, reopening :(

https://builds.apache.org/blue/organizations/jenkins/CouchDB/detail/master/231/pipeline

test/javascript/tests/users_db_security.js                     
    Error: undefined
Trace back (most recent call first):
    
  52: test/javascript/test_setup.js
      T((void 0))
  35: test/javascript/tests/users_db_security.js
      ("jchris")
  39: test/javascript/tests/users_db_security.js
      ([object Object],"org.couchdb.user:jchris","jchris")
 111: test/javascript/tests/users_db_security.js
      ("bcrypt",(function (derived_key) {TEquals(60, derived_key.length, "de
 413: test/javascript/tests/users_db_security.js
      ()
 395: test/javascript/couch_test_runner.js
      run_on_modified_server([object Array],(function () {try {testFun(schem
 411: test/javascript/tests/users_db_security.js
      ("bcrypt",1,[object Array])
 399: test/javascript/tests/users_db_security.js
      ()
  37: test/javascript/cli_runner.js
      runTest()
  48: test/javascript/cli_runner.js
      
fail

https://couchdb-vm2.apache.org/ci_errorlogs/jenkins-couchdb-231-2018-03-28T17:21:43.727450/couchlog.tar.gz

on it

Two points:

1. we are in dire need of rewriting this test file
2. it might be that this failure points to pbkdf not working as we expect it to for config admins. I’ll keep investigating.

1. Totally agree, this test and the whole javascript suite are really hard to follow. It'd be amazing to revive Russel's suggestion to migrate it to Elixir.

2. I think it's not that pbkdf or bcrypt don't work as expected, but rather hashing on heavily restricted CI hosts is not able to complete within 5 sec

In the logs from the failed jenkins' test I found:

[notice] 2018-03-28T17:21:33.282844Z [email protected] <0.32515.3> a82be2eaf1 127.0.0.1:15984 127.0.0.1 undefined PUT /_users/org.couchdb.user%3Ajchris 201 ok 511
[warning] 2018-03-28T17:21:38.285265Z [email protected] <0.32515.3> 59932a5bfc couch_httpd_auth: Authentication failed for user jchris from 127.0.0.1
[notice] 2018-03-28T17:21:38.285724Z [email protected] <0.32515.3> 59932a5bfc 127.0.0.1:15984 127.0.0.1 undefined POST /_session 401 ok 2

And when I insert something like

    if (scheme == "bcrypt") {
      wait(500);
    }
    else {
      wait(5000);
    }

at line 110 of users_db_security.js I'm getting the same failure and stack trace.

s/pbkdf/bcrypt/ apologies

This is it, we need to send log_rounds to 1 as well

Fixed in #1258