Fresheh html BUGZ:55466

* Fix typo in error message

mod_proxy(es): Avoid error response/document handling by the core if some input filter already did it while reading client's payload. When an input filter returns AP_FILTER_ERROR, it has already called ap_die() or at least already responded to the client. Here we don't want to lose AP_FILTER_ERROR when returning from proxy handlers, so we use ap_map_http_request_error() to forward any AP_FILTER_ERROR to ap_die() which knows whether a response needs to be completed or not. Before this commit, returning an HTTP error code in this case caused a double response to be generated. Depends on r1657881 to preserve r->status (for logging) when nothing is to be done by ap_die() when handling AP_FILTER_ERROR.

mod_proxy: Don't put non balancer-member workers in error state by default for connection or 500/503 errors, and honor status=+I for any error. PR 48388.

Concat string at compile time when possible. Doing so, sometimes also give the opportunity to turn a 'ap_fputstrs' into a 'ap_fputs'. PR 53741

* Correctly escape user provided data. PR: 56532 Submitted by: Maksymilian <max cert.cx> Reviewed by: rpluem

Put the uds path in its own field, and adjust the logic to look for an empty string rather than a flag.

Ensure that userland format of UDS is the same as how it is configured, no matter how we store and use it internally.

syntax sugar... if the worker is associated w/ a UDS, then make sure the log reporting has a visual clue.

Reformat the UDS support inline with a new naming structure. Use a flag for speed for testing.

More cool info ;)

Add note to balancer manager whether changes will/will not survive a restart

We are no longer pre-grabbing.

Ensure balancer runs after proxy for mutex creation

Mod_proxy used the global pool w/o mutex. fix.

Add failontimeout to allow server admin to mark balancer member in err if IO timeout occurs.

Allow for searching w/i shm slots for a specific worker and balancer

mod_proxy_balancer: It's totally unclear what Drn, Dis, Ign, Stby means. PR 52478

Use ap_rputs instead of ap_rvputs where applicable.

CVE-2012-3499 and CVE-2012-4558 Be sure to escape potential troubled strings

Allow for setting of sticky session split char... Bugz 53893

and allow the xml param

Bring XML output up to snuff for balancer-manager

fix clang warning (dead initialization)

Re-instate balancer merge (for the time being) but force persistence if the top-level server has it defined....

Use identifying server_rec info when we know we have unique and useful data :)

Persist local balancer-manager changes across restart/graceful.

No rv here.

mod_proxy_balancer: Restore balancing after a failed worker has recovered when using lbmethod_bybusyness. PR: 48735 Markus Stoll and Adam C both submitted patches against 2.2.x to bug 48735. Compared with those two, this solution 1. resets the busy field in the error-ed worker at the end of the request instead of at recovery time 2. leaves the lbstatus field alone 3. covers all possible scenarios where the busy field in the error-ed worker needs to be adjusted, since a cleanup to perform the decrement is registered at the point of the increment

* Add the forcerecovery balancer parameter that determines if recovery for balancer workers without considering the retry value of workers is enforced. There might be cases where an already overloaded backend can get into deeper trouble if the recovery of all workers is enforced without considering the retry parameter of each worker

Stray/confusing "recovery" in this failonstatus=XXX error message seems to be a copy/paste error.

Fix PR52402. Check if the shared memory was already created for this proxy_server_conf

oops, this got forgotten when copy/paste from dev branch

C/C++ 101 on how to traverse the array of pointers ;). Fix core

Various code cleanup to avoid compiler, cppcheck, or clang warnings: modules/debugging/mod_firehose.c: Make some internal functions static (to do: logs_cleanup() is unused) modules/filters/mod_charset_lite.c: Remove dead assignments modules/filters/mod_include.c: likewise modules/metadata/mod_usertrack.c: likewise modules/proxy/mod_proxy_ftp.c: likewise modules/ssl/ssl_engine_pphrase.c: likewise modules/proxy/mod_proxy_balancer.c: likewise; Remove NULL check that can never happen modules/proxy/proxy_util.c: Axe NULL-check that can never happen and if it would, it would just mask another bug os/unix/unixd.c: likewise modules/http/http_filters.c: Remove sub-condition that is always true modules/lua/mod_lua.c: Add default cases to switch statements modules/generators/mod_autoindex.c: Unsigned value can never be < 0 server/util_expr_eval.c: Fix compiler warnings with VC and on OS2

Add some more log message tags Remove some log message tags from ap_log_* calls that log lots of different error messages, in particular the config parsing errors. Not sure how we should handle those. ssl_util.c: Downgrade some dynamic locking messages from level DEBUG to TRACE1-3

Change the provider names to match the module names: mod_slotmem_shm: shared -> shm mod_socache_memcache: mc -> memcache

Add lots of unique tags to error log messages

mod_proxy: Make ap_proxy_retry_worker() into an optional function. Allows mod_lbmethod_bybusyness, mod_lbmethod_byrequests, mod_lbmethod_bytraffic and mod_lbmethod_heartbeat to be loaded without mod_proxy yet being present, which happens when modules are loaded in alphabetical order.

use a style sheet... ideally we would pull one off the filesystem, but for now...

Logic which makes the balancer "active" or not... so we can turn them off and on as needed.

Expose the path that the balancer handles...

Move useful and shared balancer stuff to the shm slot... Provide hooks for post-config/runtime addition of balancers.

Allow compile time changes of field sizes and be more "clear" regarding the diff between balancer and worker slots...

great proxy logging cleanup: * remove "proxy:", "FCGI", etc. prefixes and pid which are now included in the error log format * propagate frontend request's logconfig to backend request * use ap_log_rerror where possible * remove obsolete APLOG_NOERRNO

Cleanup effort in prep for GA push: Trim trailing whitespace... no func change

fix def'n of AP_SLOTMEM_PROVIDER_VERSION, use it

Add in BalancerGrowth and its functionality... this lays the framework for adding additional Balancers: post-config by allowing for shm growth.

Cleanup... most don't need apr_hooks.h at all...

better error message

Make the 'need to reset' more generic by extending its reach

Better knowledge of when to adjust. More info from b-m

If we change the LB method, or we take add in a new balancermember ("enable" her), then we need to have the LBmethod reset its params as required...

fix some grammar mistakes, mostly in comments

protect one-off boundary

Be at least somewhat more RESTful... Use POST for changing stuff.

No longer depend on how fork() works when laying out segments... function pointers in shm are... nasty ;)

Use ap_state_query() to fix many modules that were not correctly initializing if they were not active during server startup but got enabled later during a graceful restart (in which case they need to do all work during a single config run).

move function...

Ugg... unbalanced workers *do* need a mutex as well (good catch Rüdiger!)

typo

Squash error/warning about signed/unsigned comparison

Remove the thread mutex from the worker... it really should be in the balancer. Thus we have global and thread for the balancer. Use global when updating the full, shm list of workers; use thread when being local.

Keep track of how many workers we've used so far... no need to provide field to add another if no slots exist.

Mutex around adding a new worker...

Lock around the time when we're mucking w/ balancers...

Here we go... we can now, via balancer-manager, add new workers to existing balancers. Still work to be done, like error checking that we aren't trying to add more than we can (right now, it fails, but it would be nice to handle it nicer), disabling and *deleting* workers we don't want anymore, the actual drain method, etc... but this is some major goodness.

might as well allow for the sticky session to be adjusted as well...

And yet more balancer params that can be changed at runtime via the b-m application... next up, of course, is adding new workers ;)

Optimize... create a struct to hold the worker status info (need to align w/ set_params later on) and note when we update timestamps

Instead of just enabling/disabling, we can now adjust more useful worker status bits, such as whether hot stby, etc...

Again, abstract out the worker status magic to it's handled by funcs...

We can now change lbmethod at runtime. Provide a stub for other balancer-related params to be adjusted as well, like timeout, sticky, etc...

Move the lbmethod to also allow it to be adjusted runtime

Move some "nice to be able to change" balancer stuff to shm

Make it clear that '_' is a possible char and accepted

Make the balancer shared memory slot more human readable. But we need to make it a "safe" filename, so create 2 funcs which do that.

Align proxy_worker_shared in slotmem shm using default APR alignment.

mod_slotmem_shm is a predecessor of mod_proxy_balancer, not mod_proxy.

Display number of members possible per balancer

And check the nonce, taking care to ensure that the form actually specified a balancer.

Move nonce to balancer struct and add logic for nonce and growth. Adjust mmn and update doccos

whitespace

Builds and passes tests as prefork... good save state

Adjust for conf->workers being *proxy_worker and balancer->workers being **proxy_worker

More movement to shared stuff...

OK... a good commit point (we don't quite compile yet though...) Pull out the worker scoreboard cruft and start moving most worker stuff to shm. Use slotmem for workers and provide space for growth. Redo logic: ap_proxy_define_* ap_proxy_create_* ap_proxy_initialize_* Right now just for workers, but lay framework for balancers as well. The idea is to break out the functional parts to make it easy for dynamics. Defining is simply describing the worker, and tucking that info away. When we create, we go ahead and create the shared memory, etc... Initialize is simply to allow child process to access the shm...

Use branch now

a pointer optim.

Avoid unnecessariy initialisation before we test handler name.

mod_proxy: Rename erroronstatus to failonstatus... ... by popular demand, and add documentation. Submitted by: Daniel Ruggeri <DRuggeri primary.net>

Fold in some stuff to allow for matching of workers to their actual scoreboard slots, independent of slot ID.

remove more useless use of strlen

follow up to r930125 change log message when forcing worker into error state: * APLOG_ERR instead of unsuppressible APLOG_NOTICE * follow formatting of existing proxy messages more closely

- Be less verbose at levels INFO and DEBUG in mod_proxy* and mod_ssl - Add some trace logging to core and http

Use the new APLOG_USE_MODULE/AP_DECLARE_MODULE macros everywhere to take advantage of per-module loglevels

Move back to a post_config hook...

Add updatelbstatus to allow to have all the LB logic in the balancers.

Add in BalancerNonce directive... useful for shared-secrets.

small style tweaks to r930125

Proxy balancer: support setting error status according to HTTP response code from a backend. PR 48939 [Daniel Ruggeri <DRuggeri primary.net>]

Fix crash in balancer-manager. This was due to an incomplete refactoring in r771940 (switching balancer->workers to an array of pointers). It's likely that other balancer functionality was broken too.

In the case where we have no members, they aren't in error states :)

unused var

standard method

Allow the load balancing method to "create" workers.

Arrange the proxy_balancer_method: - Put context back after the finder(). - Use server_rec as second parameter for reset() and age(). - Add the call for reset().

update comment

Change the order of mod_proxy and mod_proxy_balancer child_init(). Change the balancer workers area to the address of workers instead copying the workers. Arrange lbmethod accordingly. Move the creation of conf->forward worker to mod_proxy child_init().

* Try first if we find any "recoverable" workers before we force them to recover.

All lbmethods are now ind submodules...

Allow determination of whether to use ';' as a sticky session path delim/sep (ala mod_jk) to be runtime (and balancer-wise) configurable.

Add in additional logging and use correct formats

Remove trailing path element, ala ./jk/native/common/jk_lb_worker.c, but use apr_strtok niceness. PR: 45158

Add in bybusyness LB method, via Joel Gluth in https://issues.apache.org/bugzilla/show_bug.cgi?id=45501 NOTE: Name changed and added in to ease enhancement/bugfixes

* Move nonce field inside the html form. PR: 45578

No functional change; simplify the CVE-2007-6420 fix slightly, as suggested by rpluem: * modules/proxy/mod_proxy_balancer.c (balancer_init): Serialize the UUID to a string here... (balancer_handler): ...rather than here.

Prevent CSRF attacks against the balancer-manager (CVE-2007-6420) * modules/proxy/mod_proxy_balancer.c (balancer_init): New function. (balancer_handler): Place a nonce in the form output, and check that the submitted form data includes that nonce. (ap_proxy_balancer_register_hook): Register the new post_config hook.

* Set at init time and combine comments.

* Do not add the query string again in the case that we are using the unparsed uri. PR: 44803

Remove all references to CORE_PRIVATE.

* Correctly escape the worker route and the worker redirect string in the HTML output of the balancer manager. Reported by SecurityReason.

* Prevent crash in balancer manager if invalid balancer name is passed as parameter.

* Set character set for HTML outputs generated by mod_ldap, mod_proxy_balancer, mod_proxy_ftp, mod_info, mod_dav without a character set to ISO-8859-1. Submitted by: jorton

* Remove unneeded include of unistd.h. getpid is not used any longer since r394065.

These really are block specific, so why not make them as such :)

* Do not reset lbstatus, lbfactor and lbset if the shared proxy_worker_stat structure for a worker is already initialized by the same or another process.

as not everyone needs sticky sessions - avoid segfault in balancer manager and mod_status

Return 503 only if the route matches some of the balancer members and nofailover=On is specified.

Fill notes "session-sticky" and subprocess_env "BALANCER_SESSION_STICKY" correctly.

Don't display "PHPSESSIONID | PHPSESSIONID" but only "PHPSESSIONID".

Remove the balancer modification logic of balancer-manager because it can't work. And add the sticky_path to the sticky display.

Add sticky_path to solve PR41897.

* Preset the form fields for route and redirect with the values of the active configuration in the scoreboard and not with the ones loaded from the configuration files.

Fix compile warning: mod_proxy_balancer.c:537: warning: unused variable 'rv'

Failure to unlock is very nasty, so log it to help with troubleshooting.

We are doing nothing here... but keep the stubs in case we ever do.

Log when the lock fails.

No need for each ind lb method to increment the elected element. Do so from the main calling func.

Add force recovery for balancer. In case all balancer members were in error state 503 was returned until the recovery timeout expired. The patch forces recovery in case all balancer members are in error state regardless of recovery timeout directive. This fixes the time gap when 503 was returned and backend was already up and running.

* Set the new environment variable BALANCER_ROUTE_CHANGED if a worker with a route different from the one supplied by the client had been chosen or if the client supplied no routing information for a balancer with sticky sessions.

Replace ap_get_server_version with ap_get_server_banner() and ap_get_server_description(). High-level summary: The full server version information is now included in the error log at startup as well as server status reports, irrespective of the setting of the ServerTokens directive. Third-party modules must now use ap_get_server_banner() or ap_get_server_description() in place of ap_get_server_version().

* Extract stickysession information correctly in the case that it is given as parameter like in the following example: www.someplace.com/somewhere/?jsessionid=gggfgdufdfoef.server&parameter=value So far we were only able to handle stickysession information encoded in the following form: www.someplace.com/somewhere/;jsessionid=gggfgdufdfoef.server?parameter=value PR: 40400

Initialization change. Move to a different format to force resets as well as a common technique, in case _route is updated at some point.

Reset standby flags for each loop through the cluster sets

Good catch by Ruediger

Add in a very simple balancer "set" concept, which allows for members to be assigned to a particular cluster set such that members in lower-numbered sets are checked/used before those in higher ones. Also bundled in this are some HTML cleanups for the balancer manager UI. Sorry for the mixins :) Compiles/builds clean: passes test framework as well as more normal usage tests ;)

The number of times elected as well as the i/o of the workers is useful info to show in the manager.

It never fails. I sit on a patch for awhile and it's not until almost right after I commit it that I think "hey, there's a better way to do that." Anyway, I was never happy about the code duplication of the primary/standby checks... This fixes that.

Add in hot-standby balancer member. If all other members are disabled or not-usable, ONLY THEN will the hot standby's be used.

update license header text

Restructure the find_route_worker() function. Basically, it should take care of knowing about usable and unusable workers. By centralizing this logic, it will make it easier and more streamlined to add in the forthcoming hot-standby status workers.

* Retry worker chosen by client supplied route / redirect worker if it is in error state before sending "Service Temporarily Unavailable". PR: 38962 Submitted by: Christian Boitel <cboitel lfdj.com> Reviewed by: rpluem

* Add the following environment variables to expose the information * about the route, the sticky session and the worker used during a request to other modules: BALANCER_SESSION_STICKY BALANCER_SESSION_ROUTE BALANCER_NAME BALANCER_WORKER_NAME BALANCER_WORKER_ROUTE PR: 39806 Submitted by: Brian <brectanu gmail.com> Reviewed by: rpluem

Merge the fcgi-proxy-dev branch to trunk, adding a FastCGI back end for mod_proxy. This log message is just a summary of the changes, for the full original log messages see r357431:393955 in branches/fcgi-proxy-dev. * modules/proxy/mod_proxy_fcgi.c: New file, holds the impementation of our new fcgi backend for mod_proxy. * modules/proxy/fcgi_protocol.h: New file, holds constants and structures for the fcgi protocol. * modules/proxy/mod_proxy_balancer.c (proxy_balancer_canon): Set up r->path_info, so the PATH_INFO env variable is correctly passed on to balancer workers. * modules/proxy/config.m4: Build the new mod_proxy_fcgi module. * support: Add fcgistarter to svn:ignore. * support/Makefile.in: Build the new fcgistarter program. * support/fcgistarter.c: New program, a helper for starting fcgi worker processes.

Update the copyright year in all .c, .h and .xml files

* Also initialize local data structures for workers of a balancer. PR: 38227, 39267 Submitted by: James A. Robinson <jim.robinson stanford.edu> Reviewed by: rpluem

* After r377780 the condition in the if clause never becomes true, because ap_proxy_initialize_worker_share is always left with PROXY_WORKER_INITIALIZED set in workers->s->status. So remove this code to save cycles and avoid confusion.

* modules/proxy/mod_proxy_balancer.c: Include unistd.h to pick up getpid() prototype.

* Do not overwrite the status of initialized workers and respect the configured status of uninitilized workers when creating a new child process.

* Fix PR37753. Off by one bug in get_path_param eats up too much of the string such that session identifier for sticky sessions is never found in URL. Submitted by: Kazuhiro Osawa <ko yappo.ne.jp> Reviewed by: Ruediger Pluem

No functional Change: Removing trailing whitespace. This also means that "blank" lines consisting of just spaces or tabs are now really blank lines

mod_proxy_balancer: BalancerManager and proxies correctly handle member workers with paths. PR36816. [Ruediger Pluem, Jim Jagielski]

* Fix a declaration problem introduced in r195013 that is C99 but not ANSI C compatible. Thus move declaration of *worker to the top of find_session_route to make it ANSI C compatible again. Submitted by: NormW <normw bocnet.com.au> Reviewed by: Ruediger Pluem

* Fix PR36507 (mod_proxy_balancer does not handle sticky sessions with tomcat correctly).

* modules/proxy/mod_proxy_balancer.c (balancer_handler): Fix gcc warning; no functional change. * modules/proxy/mod_proxy.c (set_balancer_param): Likewise.

Proxy balancer lbmethods are now registered as providers and not via hooks. Move various find() functions back to mod_proxy_balancer where they belong :)

Restruct of lbmethod find-best algo's, bypassing hook mechanism. Looking into provider method, but until then... Prevent core dump when balancer not in vhosts

Some simple debugging log entries

Arghhh.... missed the version. Sorry for the noise. Revert the hack for setting the lbmethod. The problem is deeper inside caused by the new hook to the balancer module that is never called on WIN32. Although the hack fixed core dump, the proxy is still unusable because the configuration is failing due to the empty lbmethod names array.

Revert the hack for setting the lbmethod. The problem is deeper inside caused by the new hook to the balancer module that is never called on WIN32. Although the hack fixed core dump, the proxy is still unusable because the configuration is failing due to the empty lbmethod names array.

Temporary make balancer working. What is needed is a setup for the finder function introduced, but never implemented.

When using the LB Balancer manager, if you disabled a member, then clicked on another member and/or itself again, the member would be re-enabled. This was due to the form being "submitted" and 'dw' being null (observed with Safari). Anyway, moving to a radio makes it explicit.

simple, functional interface to add additional balancer lb selection methods without requiring code changes to mod_proxy/mod_proxy_balancer; these can be implemented via sub-modules now. Let the games begin...

* modules/proxy/mod_proxy_balancer.c (proxy_balancer_pre_request): Fix another valid variable-may-be-used-uninitialized warning from GCC 4 (in the !balancer->sticky case, find_session_route returns NULL without setting route).

Set default number of maxattempts to the number of balancer members.

Simplify canon_handler. Blancer can have only a single scheme.

Add canon_handler for load balancer so that we have query string passed to remote.

Update copyright year to 2005 and standardize on current copyright owner line.

Rename proxy modules. * modules\proxy\mod_proxy_ajp.c * modules\proxy\mod_proxy_balancer.c * modules\proxy\mod_proxy_connect.c * modules\proxy\mod_proxy_ftp.c * modules\proxy\mod_proxy_http.c Renamed from proxy_{ajp,balancer,connect,ftp,http}.c * modules\proxy\mod_proxy_ajp.dsp * modules\proxy\mod_proxy_balancer.dsp * modules\proxy\mod_proxy_connect.dsp * modules\proxy\mod_proxy_ftp.dsp * modules\proxy\mod_proxy_http.dsp Update references to renamed files. * modules\proxy\NWGNUproxyajp * modules\proxy\NWGNUproxybalancer * modules\proxy\NWGNUproxycon * modules\proxy\NWGNUproxyftp * modules\proxy\NWGNUproxyhtp Update references to renamed files.