Force early setproctitle initialization by calling getproctitle()
immediately after import. This ensures setproctitle captures the
argv/environ memory layout before systemd.listen_fds() modifies
the environment by removing LISTEN_FDS and LISTEN_PID.
Without this fix, if LISTEN_FDS is the first environment variable,
setproctitle fails to detect argv correctly and silently fails.
Fixes#3430
setproctitle causes segfaults on macOS due to fork() safety issues
introduced in newer macOS versions. The mere import of setproctitle
can trigger crashes in forked worker processes.
Fixes#3021
since 3.3: EnvironmentError, IOError, socket.error and select.error are merged into IOError.
They may now return a more specific subclass - which this commit does not utilize yet.
pkg_resources is deprecated. Use the corresponding importlib.metadata
interface instead. Use the stdlib version on python >= 3.8 and use the
importlib_metadata backport on older versions.
that is, to close(0) and open /dev/null as fd=0 instead of fd=3.
(Partially) Revert "Ensure fd 0 stdin </dev/null is always inheritable."
This partially reverts commit 7946678f271e25473618929d6f2725c8c375563e.
When gunicorn --daemon daemonizes the process, prior to this change it was
noted that in the general case (without -R / --enable-stdio-inheritance), when fd 0
was replaced with /dev/null, the dup2 copy is skipped, and per PEP 446
"Make newly created file descriptors non-inheritable", the result was a stdio
fd </dev/null which was non-inheritable. As a result, any launched subprocess
did not have an open 0/stdin fd, which can cause problems in some applications.
This change retains the behaviour of opening /dev/null with fd 0, but adds a call
to os.set_inheritable(..) to ensure the fd is inheritable.
The -R branch had different logic but has now been standardised with the general
case. It was previously opening /dev/null as fd 3 and the dup2() copy made it
inheritable as fd 0. This branch now applies the same logic: open as fd 0
(i.e. after close(0)), then set_inheritable. As a result, an extra fd 3 </dev/null
previously left open is no longer left open.
Signed-off-by: Brett Randall <javabrett@gmail.com>
find_library('c') doesn't work in Alpine Linux. This happen because musl has a simpler implementation of libc.
This patch fix it by extending ctypes.util.find_library to search the libs using LD_LIBRARY_PATH.
Patch is based on e3f67780aa
See also https://bugs.python.org/issue21622fix#2160
* load application from factory function
Use `ast.parse` to validate that the string passed to the CLI is either
an attribute name or a function call. Use `ast.literal_eval` to parse
any positional and keyword arguments to the function. Call the function
to get the real application.
Co-authored-by: Connor Brinton <connor.brinton@gmail.com>
* test coverage for util.import_app
* document app factory pattern
If possible use a function based on inspect.signature to calculate the
arity. If inspect.signature is not available fall back to getargspec. The
version based on inspect.signature looks like more code but is actually a
subset of what inspect.getfullargspec does.
html.escape is preferred over cgi.escape primarily because it defaults quote
to True. For this reason pass quote=True to cgi.escape when used.
https://pypi.python.org/pypi/vulture
In particular the removal of `get_maxfd()` means the `resource` module
is no longer required (which is not available on Windows) and so helps
with #524.
This changes improve the binary upgrade behaviour using USR2:
- only one binary upgrade can happen at a time: the old arbiter needs to be
killed to promote the new arbiter.
- if a new arbiter is already spawned, until one is killed USR2 has no action
- if a new arbiter has been spawned, the unix socket won't be unlinked
- until the old arbiter have been killed the newly created pidfile has the name
<pidfile>.2 and the name Master.2 .
Note: there is no dialog between both arbiters to handle this features.
Instead they will supervise each others until one is killed. So isolation is
still guaranted.
fix#1267
This change add proper file locking to gunicorn. By default "gunicorn.lock" is created in the temporary directory when a unix socket is bound. In case someone want to fix the lock file path or use multiple gunicorn instance the "--lock-file" setting can be used to set the path of this file.
fix#1259
