http://blog.gmane.org/gmane.comp.web.lighttpd hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://comments.gmane.org/gmane.comp.web.lighttpd/5111 Stefan Bühler 2008-09-03T14:47:16 http://comments.gmane.org/gmane.comp.web.lighttpd/5109 hi there, - Centos 5.2 - Lighttpd + mod_cache built from lighttpd-1.4.19.modcache.v.1.4.4.tar.gz < at > linux.cn I have lighty running as proxy in front of some REST web services. I want to provide caching too, but mod_cache is only caching the static components (images, html, css,etc), but none of the pages with query parameters. My relevant config is : [...] cache.enable = "enable" cache.debug = "enable" cache.support-queries = "disable" cache.bases = ("/CachedResponses/") cache.refresh-pattern = ( "." => "30 update-on-refresh" ) proxy.balance = "hash" proxy.server = ( "/" => ( ( "host" => "127.0.0.1" , "port" => 81 ) ) ) proxy.worked-with-mod-cache = "enable" [...] The urls I want to cache are of the form /uriX/?param=value&param2=value2 If i enable support-queries, then all calls to /uriX/ point to the same cached result, regardless of the parameters. If I disable support-queries, it never gets cached. This matches the documentation...so here I am hoping i really have messed up my config or missed something else big. Have I ? thanks, B _________________________ {Beto|Norberto|Numard} Meijome "People demand freedom of speech to make up for the freedom of thought which they avoid. " Soren Aabye Kierkegaard I speak for myself, not my employer. Contents may be hot. Slippery when wet. Reading disclaimers makes you go blind. Writing them is worse. You have been Warned. Norberto Meijome 2008-08-29T02:37:15 http://comments.gmane.org/gmane.comp.web.lighttpd/5108 My script updates some file very often (for example, 10 times per second). Updating is done by creating tmp file, and then renaming to destination file. This file is requested by http very often also (for example, 100 times per second). File size - 4000-5000 bytes. 1. server.network-backend = "linux-sendfile" Often file is delivered to client incompletely, just few first bytes, instead of complete file. 2. server.network-backend = "gthread-aio" The same as with "linux-sendfile", but also in error_log: network_gthread_aio.c.379: (error) read() returned 0 ... not good: connections.c.1405: (trace) (network-subsys sent us a fatal-error) and sometimes: network_gthread_aio.c.383: (error) read() returned 33 *instead of 182* connections.c.1405: (trace) (network-subsys sent us a fatal-error) network_gthread_aio.c.383: (error) read() returned 31 *instead of 180* connections.c.1405: (trace) (network-subsys sent us a fatal-error) network_gthread_aio.c.383: (error) read() returned 2 *instead of 3* connections.c.1405: (trace) (network-subsys sent us a fatal-error) network_gthread_aio.c.383: (error) read() returned 511 *instead of 514* connections.c.1405: (trace) (network-subsys sent us a fatal-error) As I said before, file size is not less than 4000 bytes. I'm also confused with "instead of XXX" numbers. Any ideas? Vyacheslav Chernousov 2008-08-09T16:15:29 http://comments.gmane.org/gmane.comp.web.lighttpd/5107 I don't understand some of the ordering dependencies in the conditional configuration stuff. I'm using lighttpd 1.4.19. For example, this first method doesn't work ("server.error- handler-404" is never set): while merely swapping the order of the clauses makes it work correctly depending on the incoming request: There are times when one order is much more convenient than another so it would be better if ordering wasn't such an issue (this is just one example, I've tripped over similar ordering issues several times). Is there logic behind this that isn't obvious? Perry Perry Clarke 2008-06-29T01:00:56 http://comments.gmane.org/gmane.comp.web.lighttpd/5106 Darren Garvey 2008-06-17T18:41:17 http://comments.gmane.org/gmane.comp.web.lighttpd/5101 Hi all, I'm so not good at Unix stuff - a while ago I got a copy of the /etc/init.d/lighttpd from someone on this list but I lost it. I tried modifying from another file myself. Somehow it doesn't start the back end process (Python FCGI) properly. The Python back-end somehow can not find some modules. But if I run the same command /usr/local/sbin/lighttpd -f /etc/lighttpd.conf, it works just fine. Can anyone help point out what I'm doing wrong? Below is the file. Jack L 2008-06-14T18:06:04 http://comments.gmane.org/gmane.comp.web.lighttpd/5100 Dear lighties, thanks to darix the mailing-lists have a new home now and all previous subscribers are added to the new mailing-lists: * lighttpd< at >lists.lighttpd.net * lighttpd-announce< at >lists.lighttpd.net * lighttpd-dev< at >lists.lighttpd.net The prefixes are the same, just the domain is now < at >lists.lighttpd.net instead of < at >lists.kneschke.de. Sorry, that it took so long to bring up the mailing lists again. cheers, Jan Jan Kneschke 2008-06-08T15:08:34 http://comments.gmane.org/gmane.comp.web.lighttpd/5097 Their is a security update for libpre3 and devel packages for Ubuntu Server 6.06: Version 7.4-0ubuntu0.7.04.2: * SECURITY UPDATE: stack overflow when handling long UTF8 strings. * pcre_compile.c, testdata/test{in,out}put4: upstream changes from 7.6 backported, thanks to Tomas Hoger and Florian Weimer. * References CVE-2008-0674 I compile lighttpd from source, should I be overly concerned with the previous build without this fix? Andy Wright 2008-02-22T19:25:53 http://comments.gmane.org/gmane.comp.web.lighttpd/5096 Hello guys, first of all, thanks for this wonderful piece of software! Running Lighttpd 1.5 for about half a year now, updating to HEAD now and then. A few stability issues at the beginning have been resolved after a while. We have one problem though. We try to show a custom 503 page when our application is down for maintainance, like described over here: http://johnleach.co.uk/words/archives/2007/01/10/251/ From lighttpd.conf: magnet.attract-raw-url-to = ( "/etc/lighttpd-1.5/magnet.lua" ) magnet.lua if (lighty.stat("/u/apps/ss2-rails/shared/system/maintenance.html")) then lighty.status["maintenance.hits"] = lighty.status["maintenance.hits"] + 1 lighty.content = { { filename = "/u/apps/ss2-rails/shared/system/maintenance.html" } } lighty.header["Content-Type"] = "text/html" return 503 end When using curl or or a text browser (links), the page is correctly displayed (although the default error message is appended, see http://trac.lighttpd.net/trac/ticket/1420). In "real" browsers like Firefox and Konqueror, the page just loads forever, waiting for data to arrive (?). Furthermore, I tried using "server.errorfile-prefix" with a custom 503 page in that directory. I modified my magnet script like this, so as to return the error code only: if (lighty.stat("/u/apps/ss2-rails/shared/system/maintenance.html")) then lighty.status["maintenance.hits"] = lighty.status["maintenance.hits"] + 1 --lighty.content = { { filename = "/u/apps/ss2-rails/shared/system/maintenance.html" } } --lighty.header["Content-Type"] = "text/html" return 503 Same problem then! I don't know what's happening here. Maybe this is about problems with "Keep-Alive"? Sorry for guessing so badly, I am not a HTTP expert.. Any help would be greatly appreciated! Greetings Martin Martin Honermeyer 2008-02-22T01:58:51 http://comments.gmane.org/gmane.comp.web.lighttpd/5093 Hi All, I'm not receiving response, when i call a function from a php script. The function is from a shared library, basically a php extension . I am able to send a request to that function and receive the response with cgi enabled in my lighttpd web server. But with fast-cgi,i am not getting any response from the server. I am getting 500 error instead. Following is the lighttpd error log. ====================================================== response not received, request sent: 997 on socket: unix:/tmp/php- fastcgi.socket-5 for /symbrowser/player_control.php , closing connection ====================================================== I version of lighttpd is 1.4.13 and php version is PHP 5.2.1 (cgi-fcgi) (built: Nov 28 2007 23:09:10). Please suggest a solution asap. Thanks, Senthil. Senthil 2008-02-15T14:23:19 http://comments.gmane.org/gmane.comp.web.lighttpd/5092 Hello, When I try to save the sessions on memcached with Lighttpd 1.5 doesn't run correctly, doesn't do anything in the log of memcached, and save the session files in the default folder of php5. I have configured /etc/php5/cgi/php.ini session.save_handler = Memcache session.save_path = "tcp://192.168.100.243:11211?persistent=1" When I execute "php5-cgi -m" we can see that the module of memcache loaded. My pecl's version is 2.1.2 Memcached service: I try in 1.1.13 and 1.2.2 With apache2 the same parameters in /etc/php5/apache2/php.ini run ok. Somebody know what is the solution? Thanks sir buce 2008-02-08T11:11:08 http://comments.gmane.org/gmane.comp.web.lighttpd/5089 Are there any plans on a backport of a current Lighttpd for Debian Etch? Etch's Lighttpd is missing mode_extforward. I'm currently battling with myself if I should install Lighttpd from source or hope for a backport. Stefan Scholl 2008-02-05T08:55:18 http://comments.gmane.org/gmane.comp.web.lighttpd/5080 Dear Readers, I hope you all enjoy your seasonal holidays, may they be cold like here in Germany or warm like in Buenos Aires. Over the next days we have to leave the old mailing-list server and move to a new one. lists.kneschke.de has served us for quite some time now and the mailing-list is the last service that's not located at lighttpd.net yet. All the mail-addresses will be moved over to lists.lighttpd.net, you don't have to re-subscribe. lighttpd< at >lists.kneschke.de will become lighttpd< at >lists.lighttpd.net You will get a mail from the new mailing-list when the transition is over. cheers, Jan Jan Kneschke 2007-12-22T09:53:43 http://comments.gmane.org/gmane.comp.web.lighttpd/5070 This is regarding to http://trac.lighttpd.net/trac/ticket/1427 The problem is that this config snippet: alias.url += ( "/a/" => "/home/olaf/a/" ) $HTTP["remoteip"] == "127.0.0.1" { alias.url += ( "/doc/" => "/usr/share/doc/" ) } alias.url += ( "/b/" => "/home/olaf/b/" ) Does not do what anyone not familiar with lighttpd internals would expect: the alias for "/b/" doesn't work. Moving it before the conditional works. (When accessing via localhost). The first reply to the ticket, from "moo", claims that this is intentional. If so, it's a really strange decision, and undocumented. Anyone want to comment? Jan? Regards, Steve Steve Greenland 2007-12-13T21:27:03 http://comments.gmane.org/gmane.comp.web.lighttpd/5067 Hi, I'm using lighttpd to start some Python fastcgi processes. When I killall lighttpd, the Python processes stay. Is this the expected behavior? Is "killall lighttpd" the right way to kill lighttpd? Jack L 2007-12-11T02:49:01 http://comments.gmane.org/gmane.comp.web.lighttpd/5066 hi, if i spawn an fcgi processes using spawn-fcgi, and exit the terminal it does not close till the spawned fcgi process dies or is killed in opensolaris... how do i fix this? rkmr.em< at >gmail.com 2007-12-11T01:23:11 http://comments.gmane.org/gmane.comp.web.lighttpd/5060 http://blog.davber.com/2007/12/10/web-server-performance-shoot-out-simple-pages/ Eugen Leitl 2007-12-10T17:56:35 http://comments.gmane.org/gmane.comp.web.lighttpd/5058 hello everyone, I have lighttpd 1.4.18-1, running on Linux RHEL4 (with updates from centos4). I set it up to run as a proxy for a Zope app which runs on port 8080. It works ok via port 80 and 443 - thanks for making it so easy! Now, the zope app doesn't compress it's content. I would like lighty to do it for me. So I enabled the compress module, configured it..and nothing changes. I changed the order the module are defined in the config file, hoping that would affect the order in which they are loaded/ executed, no luck. My relevant bits of configs are : ----------------------------------- server.modules = ( "mod_redirect", "mod_access", "mod_compress", "mod_proxy", "mod_accesslog" ) ## change uid to <uid> (default: don't care) server.username = "lighttpd" ## change uid to <uid> (default: don't care) server.groupname = "lighttpd" #### compress module compress.cache-dir = "/var/cache/lighttpd/compress/" compress.filetype = ("text/plain", "text/html") proxy.debug=0 proxy.server = ( "" => ( ( "host" => "127.0.0.1", "port" => 8080 ) ) ) ---------------------------- [root< at >nms lighttpd]# ls -l /var/cache/ drwxr-xr-x 3 root root 4096 Dec 10 12:56 lighttpd [root< at >nms lighttpd]# ls -l /var/cache/lighttpd/ drwxr-xr-x 2 lighttpd lighttpd 4096 Dec 10 12:56 compress the compress directory is also empty. What am I missing? thanks in advance! Beto _________________________ {Beto|Norberto|Numard} Meijome If you find a solution and become attached to it, the solution may become your next problem. I speak for myself, not my employer. Contents may be hot. Slippery when wet. Reading disclaimers makes you go blind. Writing them is worse. You have been Warned. Norberto Meijome 2007-12-10T14:19:08 http://comments.gmane.org/gmane.comp.web.lighttpd/5056 Hello, in some older releases (til 1.4.11 or even earlier) lighttpd had hardcoded request handling timeout - eg time lighttpd waits response to be generated from backend (php) after sucessfull connection. If it was exceeded the request was closed and php child freed. By default the timeout was 360 secs (or something like that (not a configurable value)) I've changed it to 30 to be inline with php's max_execution_time .. It was good because in some scripts where php uses sockets sometimes it doesnt honour its own script max execution time so webserver ends up with a lot of requests being in handle-req state (according to server-status). I can't find this in 1.4.18 anymore (or it has been moved) and my changes to those older versions are gone (by mistake). So maybe somebody can point me to the right source where something like that could be implemented/changed? wbr Reinis Reinis Rozitis 2007-12-02T15:57:08 http://comments.gmane.org/gmane.comp.web.lighttpd/5054 Hi all, I'm using a Python CGI with lighttpd. a form is submitted to this CGI. I find that if I use method POST, this line of Python code gets the form parameters correctly, but if I use GET, the form variable does not contain any parameters: form = cgi.FieldStorage(keep_blank_values = True) The same script is tested working OK on two other web servers. Am I missing something? Jack L 2007-12-01T17:54:34 http://comments.gmane.org/gmane.comp.web.lighttpd/5047 I'm using mod_proxy_core to run some fast-cgi apps, but I'm having problems with the QUERY_STRING env variable, which is not being passed to the application... Also, somewhere in the docs of fast-cgi, it is specified that if I want to change the http status, I must set a Status header, which will be read by the mod_fcgi (apache) and will change the http status... is this the expected behaviour in lighty? if so, it's also not working in 1.5. Regards, Rolando Abarca 2007-11-28T16:48:59 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.web.lighttpd