more consistent cleanup of resources at shutdown
(e.g. upon error conditions)
Notes: graceful restart with SIGUSR1
- not available if chroot()ed, oneshot mode, or if idle timeout occurs
- preserve process id (pid)
- preserve existing listen sockets
- i.e. does not close old listen sockets from prior configs
(even if old listen sockets no longer in the new config)
(sockets may have been bound w/ root privileges no longer available)
- will fail to add listen sockets from new config if privileges
lighttpd configured to drop privileges to non-root user, and
new listen socket attempts to bind to low-numbered port requiring
root privileges.
- will fail if listen sockets in new config conflict with any previous
old listen sockets
- These failure modes will result in lighttpd shutting down instead of
graceful restart. These failure modes are not detectable with
preflight checks ('lighttpd -tt -f lighttpd.conf') because the
new instance of lighttpd running the preflight check does not
known config state of n prior graceful restarts, or even the
config state of the currently running lighttpd server.
- due to lighttpd feature of optionally managing backends
(e.g. fastcgi and scgi via "bin-path"), lighttpd must wait for
all child processes to exit prior to restarting. Restarting new
workers while old workers (and old backends) were still running would
result in failure of restarted lighttpd process to be able to bind to
sockets already in use by old backends (e.g. unix "socket" path)
x-ref:
"graceful restart with SIGUSR1"
https://redmine.lighttpd.net/issues/2785
close connections in keep-alive that are waiting for next request
disable keep-alive on existing connections
remove bandwidth write limits
reduce remaining linger timeout (on already finished requests)
to be (from zero) *up to* one more second, but no more
large code move, but minimal changes made to code (besides whitespace),
so that code builds
next: need to isolate openssl data structures and config parsing
handle_request_env (called on demand by handlers to populate env)
handle_connection_accept
handle_connection_shut_wr (was handle_connection_close)
handle_connection_close (now occurs at socket close())
[mod_deflate] skip deflate if 1 min loadavg too high
deflate.max-loadavg = "3.50" # express value as string of float num
[mod_compress] skip compression if 1 min loadavg too high
compress.max-loadavg = "3.50" # express value as string of float num
Feature available on BSD-like systems which have getloadavg() in libc
Note: load average calculations are different on different operating
systems and different types of system loads, so there is no value that
can be recommended for one-size-fits-all.
x-ref:
"Enable mod_compress to abandon compression when load average is too high"
https://redmine.lighttpd.net/issues/1505
prefer RAND_pseudo_bytes() (openssl), arc4random() or jrand48(),
if available, over rand()
These are not necessarily cryptographically secure, but should be better
than rand()
limits total size per request of request headers submitted by client
default limit set to 8k (prior lighttpd <= 1.4.41 hard-coded 64k limit)
(similar to Apache directive LimitRequestFieldSize)
x-ref:
"limits the size of HTTP request header"
https://redmine.lighttpd.net/issues/2130
module status: experimental; more testing and review needed
Kerberos library calls have been preserved from original patch set
and should be reviewed.
module has been quickly tested with Basic auth (Use over TLS!)
SPNEGO -has not- been tested. Again, kerberos library calls have
been preserved from original patch set. YMMV. (Use over TLS!)
x-ref:
"Kerberos/GSSAPI Delegation Support"
https://redmine.lighttpd.net/issues/1899
server.username can not be root or 0.
server.groupname can not be root or 0.
If server.username is set, previous behavior might retain gid 0
if server.groupname was not set.
New behavior calls setgid() on server.username primary gid, and
then initgroups on server.username if server.username is set but
server.groupname is not set.
x-ref:
"server.groupname not required with server.username"
https://redmine.lighttpd.net/issues/2725
Do not switch to CON_STATE_ERROR upon idle timeout if already in
CON_STATE_CLOSE. Changing to CON_STATE_ERROR might keep resetting
con->close_timeout_ts if repeated calls to shutdown() succeed.
This allows admin to configure if response is collected in entirety
prior to sending data to client
For compatibility with existing configs, default is existing behavior:
buffer entire response prior to sending data to client
The following are config options, though not all implemented yet
// default: buffer entire request body before connecting to backend
server.stream-request-body = 0
// stream request body to backend; buffer to temp files
server.stream-request-body = 1
// stream request body to backend; minimal buffering might block upload
server.stream-request-body = 2
// default: buffer entire response body before sending to client
server.stream-request-body = 0
// stream response body to client; buffer to temp files
server.stream-request-body = 1
// stream response body to client; minimal buffering might block backend
server.stream-request-body = 2
x-ref:
"fastcgi, cgi, flush, php5 problem."
https://redmine.lighttpd.net/issues/949
"Reimplement upload (POST) handling to match apache/zeus/thttpd/boa functionality"
https://redmine.lighttpd.net/issues/376
server.http-parseopt-header-strict = "enable"
server.http-parseopt-host-strict = "enable" (implies host-normalize)
server.http-parseopt-host-normalize = "disable"
defaults retain current behavior, which is strict header parsing
and strict host parsing, with enhancement to normalize IPv4 address
and port number strings.
For lighttpd tests, these need to be enabled (and are by default)
For marginally faster HTTP header parsing for benchmarks, disable these.
To allow
- underscores in hostname
- hypen ('-') at beginning of hostname
- all-numeric TLDs
server.http-parseopt-host-strict = "disable"
x-ref:
"lighttpd doesn't allow underscores in host names"
https://redmine.lighttpd.net/issues/551
"hyphen in hostname"
https://redmine.lighttpd.net/issues/1086
"a numeric tld"
https://redmine.lighttpd.net/issues/1184
"Numeric tld's"
https://redmine.lighttpd.net/issues/2143
"Bad Request"
https://redmine.lighttpd.net/issues/2258
"400 Bad Request when using Numeric TLDs"
https://redmine.lighttpd.net/issues/2281
To allow a variety of numerical formats to be converted to IP addresses
server.http-parseopt-host-strict = "disable"
server.http-parseopt-host-normalize = "enable"
x-ref:
"URL encoding leads to "400 - Bad Request""
https://redmine.lighttpd.net/issues/946
"400 Bad Request when using IP's numeric value ("ip2long()")"
https://redmine.lighttpd.net/issues/1330
To allow most 8-bit and 7-bit chars in headers
server.http-parseopt-header-strict = "disable" (not recommended)
x-ref:
"Russian letters not alowed?"
https://redmine.lighttpd.net/issues/602
"header Content-Disposition with russian '?' (CP1251, ascii code 255) causes error"
https://redmine.lighttpd.net/issues/1016
libressl defines SSL_OP_NO_SSLv2 and SSL_OP_NO_SSLv3 as 0x0
(thx Christian Heckendorf)
libressl matches ERR_remove_thread_state() signature from openssl 1.0.2
(libressl pretends that libressl is openssl version 2.0.0,
but openssl 1.1.0 changes signature of ERR_remove_thread_state())
libressl does not yet provide compatibility interfaces for the new
prototypes introduced in openssl 1.1.0, including
DH_set0_pqg() and DH_set_length()
remove OPENSSL_NO_KRB5 from build config (added in 5fab991b in 2005)
(define USE_OPENSSL_KERBEROS if required)
(Note: OPENSSL_NO_KRB5 removed in openssl 1.1.0)
(e.g. when called from xinetd)
Note: lighttpd is designed as a high performance, long-running server,
not a one-shot executable. This one-shot mode of operation has not been
tuned for performance. lighttpd server start-up and initialization aims
for correctness, not speed. If using this one-shot mode as part of fork
and exec from xinetd, then performance is already not of high concern.
x-ref:
"support for xinetd"
https://redmine.lighttpd.net/issues/1584
-i <secs> graceful shutdown after <secs> of inactivity
Option might be used with applications such as git instaweb.
While git instaweb does have command line options of its own
to [start,stop,restart], some may find it convenient to configure
git instaweb to start lighttpd with a default idle time limit,
after which lighttpd will gracefully shut itself down without
any further action from the user.
x-ref:
"[PATCH] support -i <secs> idle timeout option"
https://redmine.lighttpd.net/issues/2696
original request and patch submitted by mackyle. thx.
server.error-handler preserves HTTP status error code when error page
is static, and allows dynamic handlers to change HTTP status code
when error page is provided by dynamic handler. server.error-handler
intercepts all HTTP status codes >= 400 except when the content is
generated by a dynamic handler (cgi, ssi, fastcgi, scgi, proxy, lua).
The request method is unconditionally changed to GET for the request
to service the error handler, and the original request method is
later restored (for logging purposes). request body from the
original request, if present, is discarded.
server.error-handler is somewhat similar to server.error-handler-404,
but server.error-handler-404 is now deprecated, intercepts only 404
and 403 HTTP status codes, and returns 200 OK for static error pages,
a source of confusion for some admins. On the other hand, the new
server.error-handler, when set, will intercept all HTTP status error
codes >= 400. server.error-handler takes precedence over
server.error-handler-404 when both are set.
NOTE: a major difference between server.error-handler and the
now-deprecated server.error-handler-404 is that the values of the
non-standard CGI environment variables REQUEST_URI and REDIRECT_URI
have been swapped. Since REDIRECT_STATUS is the original HTTP
status code, REDIRECT_URI is now the original request, and REQUEST_URI
is the current request (e.g. the URI/URL to the error handler).
The prior behavior -- which reversed REQUEST_URI and REDIRECT_URI values
from those described above -- is preserved for server.error-handler-404.
Additionally, REDIRECT_STATUS is now available to mod_magnet, which
continues to have access to request.uri and request.orig_uri.
See further discussion at https://redmine.lighttpd.net/issues/2702
and https://redmine.lighttpd.net/issues/1828
github: closes #36
remove handle_joblist hook and remove the hooks defined in
mod_fastcgi and mod_scgi. The calls made to fdevent management
are redundant. If the calls were actually needed, then
mod_proxy would have needed a handle_joblist handler, too.
to detect client disconnect. Do so even when waiting on backend,
and not polling for POLLRD or POLLWR on client connection.
This reduces unnecessary load on backends when backends are slow
to respond and client has given up waiting.
x-ref:
"https://redmine.lighttpd.net/issues/399"
FastCGI performance on high load
For backwards compatibility with existing lighttpd configs, default is
mimetype.xattr-name = "Content-Type"
Those who wish to use the freedesktop.org definition of xattr mimetype
can set the following in the global lighttpd config:
mimetype.xattr-name = "user.mime_type"
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3131 152afb58-edef-0310-8abb-c4023f1b3aa9
lighttpd -t loads config file and performs syntax check
lighttpd -tt (new) performs preflight startup checks,
including loading and initializing modules, but skipping any
potentially destructive actions which might affect an already
running server (separate instance). These currently include:
- skipping pidfile modification
- skipping bind() to network sockets
- skipping open of error and access logs
From: Glenn Strauss <gstrauss@gluelogic.com>
git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@3130 152afb58-edef-0310-8abb-c4023f1b3aa9
