| Date | Commit message (Collapse) |
|---|
|
In the case where preload_app is true, delay binding new
listeners until after loading the application.
Some applications have very long load times (especially Rails
apps with Ruby 1.9.2). Binding listeners early may cause a load
balancer to incorrectly believe the unicorn workers are ready to
serve traffic even while the app is being loaded.
Once a listener is bound, connect() requests from the load
balancer succeed until the listen backlog is filled. This
allows requests to pile up for a bit (depending on backlog size)
before getting rejected by the kernel. By the time the
application is loaded and ready-to-run, requests in the
listen backlog are likely stale and not useful to process.
Processes inheriting listeners do not suffer this effect, as the
old process should still be capable of serving new requests.
This change does not improve the situation for the
preload_app=false (default) use case. There may not be a
solution for preload_app=false users using large applications.
Fortunately Ruby 1.9.3+ improves load times of large
applications significantly over 1.9.2 so this should be less of
a problem in the future.
Reported via private email sent on 2012-06-29T22:59:10Z
|
|
It's too much overhead to keep Rails-specific tests working,
especially when it's hauling in an ancient version of SQLite3.
Since Rails 3 has settled down with Rack and unicorn_rails is
unlikely to need changing in the future, we can drop these
tests.
|
|
These should be made executable for ease-of-understanding and
consistency, regardless of whether we actually execute them.
|
|
Previously we relied on implicit socket shutdown() from the
close() syscall. However, some Rack applications fork()
(without calling exec()), creating a potentially long-lived
reference to the underlying socket in a child process. This
ends up causing nginx to wait on the socket shutdown when the
child process exits.
Calling shutdown() explicitly signals nginx (or whatever client)
that the unicorn worker is done with the socket, regardless of
the number of FD references to the underlying socket in
existence.
This was not an issue for applications which exec() since
FD_CLOEXEC is always set on the client socket.
Thanks to Patrick Wenger for discovering this. Thanks to
Hongli Lai for the tip on using shutdown() as is done in
Passenger.
ref: http://mid.gmane.org/CAOG6bOTseAPbjU5LYchODqjdF3-Ez4+M8jo-D_D2Wq0jkdc4Rw@mail.gmail.com
|
|
This seems required for TLSv1.2 under OpenSSL 1.0.1
|
|
We're only allowed 108 bytes for Unix domain sockets.
mktemp(1) usually generates path names of reasonable length
and we rely on it anyways.
|
|
The output of SHA1 command-line tools is too unstable and
I'm more comfortable with Ruby 1.9 encoding support than
I was in 2009.
Jeremy Evans noted the output of "openssl sha1" has
changed since I last used it.
|
|
expr on OpenBSD uses a basic regular expression (according to
re_format(7)), which doesn't support +, only *.
Acked-by: Eric Wong <normalperson@yhbt.net>
|
|
We throw up some fake SSL certs for testing
|
|
This will also be the foundation of SSL support in Rainbows!
and Zbatery. Some users may also want to use this in
Unicorn on LANs to meet certain security/auditing requirements.
Of course, Nightmare! (in whatever form) should also be able to
use it.
|
|
Backtraces are now formatted properly (with timestamps) and
exceptions will be logged more consistently and similar to
Logger defaults:
"#{exc.message} (#{e.class})"
backtrace.each { |line| ... }
This may break some existing monitoring scripts, but errors
will be more standardized and easier to check moving forward.
|
|
"app error" is more correct, and consistent with Rainbows!
|
|
rescuing from SystemExit and exit()-ing again is ugly, but
changes made to lower stack depth positively affect _everyone_
so we'll tolerate some ugliness here.
We'll need to disable graceful exit for some tests, too...
|
|
Just in case we break anything
|
|
Rainbows! wants to be able to lower this eventually...
|
|
"wait" needs to be done in the outside shell because
the subshell could still be exiting when we grep.
|
|
There's an HTTP status code allocated for it in
<http://www.iana.org/assignments/http-status-codes>, so
return that instead of 400.
|
|
This was broken since v3.3.1[1] since nginx relies on a closed
socket (and not Content-Length/Transfer-Encoding) to detect
a response completion. We have to close the client socket
before invoking GC to ensure the client sees the response
in a timely manner.
[1] - commit b72a86f66c722d56a6d77ed1d2779ace6ad103ed
|
|
No need to use an ancient Rack now that we've dropped Rails
2.3.x tests. We need to remember that Rack 1.1.0 doesn't
support input#size.
|
|
This causes conflicts with ports clients may use in
the ephemeral range since those do not hold FS locks.
This reverts commit e597e594ad88dc02d70f7d3521d0d3bdc23739bb.
Conflicts:
test/test_helper.rb
|
|
Duh...
|
|
Response bodies may capture the block passed to each
and save it for body.close, so don't close the socket
before we have a chance to call body.close
|
|
More config bloat, sadly this is necessary for Rainbows! :<
|
|
Since modern machines have more memory these days and
clients are sending more data, avoiding potentially slow
filesystem operations for larger uploads can be useful
for some applications.
|
|
This has been broken since 2.0.x
Internal cleanups sometimes have unintended consequences :<
|
|
In case a request sends the header and buffer as one packet,
TeeInput relying on accounting info from StreamInput is harmful
as StreamInput will buffer in memory outside of TeeInput's
control.
This bug is triggered by calling env["rack.input"].size or
env["rack.input"].rewind before to read.
|
|
This will help ensure we trap our own errors properly
in the future.
|
|
oops :x
|
|
This allows users to override the current Rack spec and disable
the rewindable input requirement. This can allow applications
to use less I/O to minimize the performance impact when
processing uploads.
|
|
We need to ensure the old worker is reaped before sending
new requests intended for the new worker.
|
|
If a configuration directive is set at startup and later
unset, it correctly restores the original default value
as if it had never been set in the first place.
This applies to the majority of the configuration values with
a few exceptions:
* This only applies to stderr_path and stdout_path when
daemonized (the usual case, they'll be redirected to
"/dev/null"). When NOT daemonized, we cannot easily redirect
back to the original stdout/stderr destinations.
* Unsetting working_directory does not restore the
original working directory where Unicorn was started.
As far as we can tell unsetting this after setting it is
rarely desirable and greatly increases the probability of
user error.
|
|
This may be useful for some apps that wish to drain the body
before acquiring an app-wide lock. Maybe it's more useful
with Rainbows!...
|
|
Automation is nice, the makefile needs some cleanup
|
|
We'll be using more of Isolate in development.
|
|
While we've always unlinked dead sockets from nuked/leftover
processes, blindly unlinking them can cause unnecessary failures
when an active process is already listening on them. We now
make a simple connect(2) check to ensure the socket is not in
use before unlinking it.
Thanks to Jordan Ritter for the detailed bug report leading to
this fix.
ref: http://mid.gmane.org/8D95A44B-A098-43BE-B532-7D74BD957F31@darkridge.com
|
|
Rails 3 is out, and requires no code changes on our end to work
(as far as our tests show :)
|
|
These nasty hacks were breaking Rubinius compatibility.
This can be further cleaned up, too.
|
|
No code changes needed, thankfully.
|
|
Something is wrong if workers exit with a non-zero status,
so we'll increase the log level to help prevent people
from missing it.
|
|
In addition to SIGHUP, it should be possible to gradually bring
workers back up (to avoid overloading the machine) when rolling
back upgrades after SIGWINCH.
Noticed-by: Lawrence Pit
ref: http://mid.gmane.org/4C3F8C9F.2090903@gmail.com
|
|
As described in our SIGNALS documentation, sending SIGHUP to the
old master (to respawn SIGWINCH-ed children) while the new
master (spawned from SIGUSR2) is active is useful for backing
out of an upgrade before sending SIGQUIT to the new master.
Unfortunately, the SIGHUP signal to the old master will cause
the ".oldbin" pid file to be reset to the non-".oldbin" version
and thus attempt to clobber the pid file in use by the
to-be-terminated new master process.
Thanks to the previous commit to prevent redaemonization in the
new master, the old master can reliably detect if the new master
is active while it is reloading the config file.
Thanks to Lawrence Pit for discovering this bug.
ref: http://mid.gmane.org/4C3BEACF.7040301@gmail.com
|
|
test-lib handles variables named "*socket" (and "*fifo")
differently than ordinary variables.
|
|
Our fugly code can't handle embedded command-line options in
config.ru when using Rubinius yet. So add some related tests
to the ones marked RBX_SKIP that don't rely on embedded
command-line options.
|
|
While log reopening worked reliably for newly-created File
objects in the unit tests, the $stderr and $stdout handles that
get redirected did not get reopened reliably under Rubinius.
We work around this by relying on Rubinius internals and
directly setting the @path instance variable. This is harmless
for MRI and should be harmless for other any other Ruby
implementations we'll eventually support.
ref: http://github.com/evanphx/rubinius/issues/360
|
|
Ruby 1.9.2 no longer adds the current directory to $LOAD_PATH
automatically.
|
|
|
|
|
|
We share the same array from the original bin/* down
into the Configurator.
|
|
This was commit abc207b2918606867094f2820bab58223e99aac4 from
rainbows.git
|
|
|