3 Occasionally odd {issues}[link:ISSUES.html] arise without a transparent or
4 acceptable solution. Those issues are documented here.
6 * Issues with FreeBSD jails can be worked around as documented by Tatsuya Ono:
7 http://mid.gmane.org/CAHBuKRj09FdxAgzsefJWotexw-7JYZGJMtgUp_dhjPz9VbKD6Q@mail.gmail.com
9 * PRNGs (pseudo-random number generators) loaded before forking
10 (e.g. "preload_app true") may need to have their internal state
11 reset in the after_fork hook. Starting with \Unicorn 3.6.1, we
12 have builtin workarounds for Kernel#rand and OpenSSL::Random users,
13 but applications may use other PRNGs.
15 * Under some versions of Ruby 1.8, it is necessary to call +srand+ in an
16 after_fork hook to get correct random number generation. We have a builtin
17 workaround for this starting with \Unicorn 3.6.1
19 See http://redmine.ruby-lang.org/issues/show/4338
21 * On Ruby 1.8 prior to Ruby 1.8.7-p248, *BSD platforms have a broken
22 stdio that causes failure for file uploads larger than 112K. Upgrade
23 your version of Ruby or continue using Unicorn 1.x/3.4.x.
25 * For notes on sandboxing tools such as Bundler or Isolate,
26 see the {Sandbox}[link:Sandbox.html] page.
28 * nginx with "sendfile on" under FreeBSD 8 is broken when
29 uploads are buffered to disk. Disabling sendfile is required to
30 work around this bug which should be fixed in newer versions of FreeBSD.
32 * When using "preload_app true", with apps using background threads
33 need to restart them in the after_fork hook because threads are never
34 shared with child processes. Additionally, any synchronization
35 primitives (Mutexes, Monitors, ConditionVariables) should be
36 reinitialized in case they are held during fork time to avoid
37 deadlocks. The core Ruby Logger class needlessly uses a MonitorMutex
38 which can be disabled with a {monkey patch}[link:examples/logger_mp_safe.rb]
42 * Under Ruby 1.9.1, methods like Array#shuffle and Array#sample will
43 segfault if called after forking. Upgrade to Ruby 1.9.2 or call
44 "Kernel.rand" in your after_fork hook to reinitialize the random
47 See http://redmine.ruby-lang.org/issues/show/2962 for more details
49 * Rails 2.3.2 bundles its own version of Rack. This may cause subtle
50 bugs when simultaneously loaded with the system-wide Rack Rubygem
51 which Unicorn depends on. Upgrading to Rails 2.3.4 (or later) is
52 strongly recommended for all Rails 2.3.x users for this (and security
53 reasons). Rails 2.2.x series (or before) did not bundle Rack and are
54 should be unnaffected. If there is any reason which forces your
55 application to use Rails 2.3.2 and you have no other choice, then
56 you may edit your Unicorn gemspec and remove the Rack dependency.
58 ref: http://mid.gmane.org/20091014221552.GA30624@dcvr.yhbt.net
59 Note: the workaround described in the article above only made
60 the issue more subtle and we didn't notice them immediately.
62 * WONTFIX: code reloading and restarts with Sinatra 0.3.x (and likely older
63 versions) apps is broken. The workaround is to force production
64 mode to disable code reloading as well as disabling "run" in your
68 Since this is no longer an issue with Sinatra 0.9.x apps, this will not be
69 fixed on our end. Since Unicorn is itself the application launcher, the
70 at_exit handler used in old Sinatra always caused Mongrel to be launched
71 whenever a Unicorn worker was about to exit.
73 Also remember we're capable of replacing the running binary without dropping
74 any connections regardless of framework :)