From tom at jinstech.com Wed Jan 12 05:34:45 2005 From: tom at jinstech.com (Tom Jin) Date: Tue, 11 Jan 2005 21:34:45 -0800 Subject: tux kernel panic Message-ID: <001901c4f868$6d0444a0$6600a8c0@ye.neet> Hi: This problem has been reported already. tux crashed with symbolic links in FC3. Here's my tests: 1. Default Fedora Core 3 installation. Symbolic links located in both reiserfs and ext3 partitions, kernel panic. 2. FC3 with kernel 2.6.10 and tux3-2.6.10-rc3-A7. Same panic. Some of dmesgs: Jan 11 20:22:23 img01 kernel: ------------[ cut here ]------------ Jan 11 20:22:23 img01 kernel: kernel BUG at fs/namei.c:497! Jan 11 20:22:23 img01 kernel: invalid operand: 0000 [#21] Jan 11 20:22:23 img01 kernel: CPU: 0 Jan 11 20:22:23 img01 kernel: EIP: 0060:[] Not tainted VLI Jan 11 20:22:23 img01 kernel: EFLAGS: 00010206 (2.6.10) Jan 11 20:22:23 img01 kernel: EIP is at link_path_walk+0x70f/0xcef Jan 11 20:22:23 img01 kernel: eax: f7cfda60 ebx: f6e509b7 ecx: f7408e48 edx: f7e2434c Jan 11 20:22:23 img01 kernel: esi: f7408000 edi: f7408e0c ebp: f73a030c esp: f7408dd4 Jan 11 20:22:23 img01 kernel: ds: 007b es: 007b ss: 0068 Jan 11 20:22:23 img01 kernel: Process tux (pid: 3344, threadinfo=f7408000 task=f7cfda60) Jan 11 20:22:23 img01 kernel: Stack: f7408e04 f7408e08 f7408e04 00000040 00000001 f7408e28 c0121d07 ffffffd8 Jan 11 20:22:23 img01 kernel: 00000001 00000040 f7e2434c f7408e48 c20f4a00 f73a030c da15a576 00000006 Jan 11 20:22:23 img01 kernel: f6e509b0 00000000 f7408e48 00000040 00000040 c0383028 c029ec7e f6e50800 Jan 11 20:22:23 img01 kernel: Call Trace: Jan 11 20:22:23 img01 kernel: [] autoremove_wake_function+0x29/0x4b Jan 11 20:22:23 img01 kernel: [] __tux_lookup+0x10/0x21 Jan 11 20:22:23 img01 kernel: [] tux_lookup+0xba/0xf8 Jan 11 20:22:23 img01 kernel: [] release_console_sem+0x93/0x98 Jan 11 20:22:23 img01 kernel: [] printk+0x17/0x1b Jan 11 20:22:23 img01 kernel: [] lookup_url+0x97/0x620 Jan 11 20:22:23 img01 kernel: [] parse_http_message+0x1b29/0x206d Jan 11 20:22:23 img01 kernel: [] http_process_message+0x0/0x2db Jan 11 20:22:23 img01 kernel: [] release_sock+0x3e/0x6c Jan 11 20:22:23 img01 kernel: [] http_process_message+0x53/0x2db Jan 11 20:22:23 img01 kernel: [] process_requests+0xa0/0xb5 Jan 11 20:22:23 img01 kernel: [] event_loop+0x112/0x1a9 Jan 11 20:22:23 img01 kernel: [] __sys_tux+0x82/0x7e2 Jan 11 20:22:23 img01 kernel: [] sys_chdir+0x6f/0x77 Jan 11 20:22:23 img01 kernel: [] sys_setuid+0xb5/0xc1 Jan 11 20:22:23 img01 kernel: [] sysenter_past_esp+0x52/0x75 Jan 11 20:22:23 img01 kernel: Code: 54 24 2c 89 3c 24 89 54 24 04 e8 63 f7 ff ff 8b 4c 24 2c 8b 11 8b 52 08 89 54 24 28 e9 4f f9 ff ff e8 ca 54 17 00 e9 3f fb ff ff <0f> 0b f1 01 6a 1f 2d c0 90 e9 26 fb ff ff 83 4c 24 20 03 8b 4c Jan 11 20:22:23 img01 kernel: <5>TUX: thread 0 stopping ... -------------- next part -------------- An HTML attachment was scrubbed... URL: From mcd at daviesinc.com Wed Jan 12 06:56:04 2005 From: mcd at daviesinc.com (Chris Davies) Date: Wed, 12 Jan 2005 01:56:04 -0500 Subject: tux kernel panic In-Reply-To: <001901c4f868$6d0444a0$6600a8c0@ye.neet> References: <001901c4f868$6d0444a0$6600a8c0@ye.neet> Message-ID: <1105512964.3479.3.camel@tsavo> I know we ran into this using Debian/3.0 If I recall, a symlink on the 'domain' directory, i.e. /var/www/domain.com -> /home/httpd/username/domain.com will cause that, but, /var/www/domain.com/index.html -> /var/www/domain.com/index.whatever will work. If I recall, it was only when the symlink was the root of the domain. On Tue, 2005-01-11 at 21:34 -0800, Tom Jin wrote: > Hi: > > This problem has been reported already. > tux crashed with symbolic links in FC3. > From mail at a-wing.co.uk Sun Jan 16 09:14:19 2005 From: mail at a-wing.co.uk (Andrew Hutchings) Date: Sun, 16 Jan 2005 09:14:19 +0000 Subject: Virtual Hosting Message-ID: <41EA306B.2020305@a-wing.co.uk> Is there any way to rediect the request http://www.domain.com to $DOCROOT/domain.com/www Or is it something I should write a module / patch for? Regards Andrew -- Andrew Hutchings Systems Operator / Developer / NWLUG Admin (http://www.nwlug.org.uk) Netserve Consultants Ltd. http://www.domaincity.co.uk/ From sgr at logsoft.com Sun Jan 16 16:53:14 2005 From: sgr at logsoft.com (Stephen Reppucci) Date: Sun, 16 Jan 2005 11:53:14 -0500 (EST) Subject: Virtual Hosting In-Reply-To: <41EA306B.2020305@a-wing.co.uk> Message-ID: Check out mod_vhost_alias. I use this extensively, and it does a wonderful job. On Sun, 16 Jan 2005, Andrew Hutchings wrote: > Is there any way to rediect the request http://www.domain.com to > $DOCROOT/domain.com/www > Or is it something I should write a module / patch for? > > Regards > Andrew > -- Steve Reppucci sgr at logsoft.com | Logical Choice Software http://logsoft.com/ | =-=-=-=-=-=-=-=-=-=- My God! What have I done? -=-=-=-=-=-=-=-=-=-= From mail at a-wing.co.uk Sun Jan 16 22:06:56 2005 From: mail at a-wing.co.uk (Andrew Hutchings) Date: Sun, 16 Jan 2005 22:06:56 +0000 Subject: Virtual Hosting In-Reply-To: References: Message-ID: <41EAE580.8030708@a-wing.co.uk> Hi Stephen, That is what we are using to do it in Apache (worked very well for many years), I didn't know it was written for Tux as well, where can I get a Tuxified version? Regards Andrew Stephen Reppucci wrote: > Check out mod_vhost_alias. I use this extensively, and it does a > wonderful job. > > On Sun, 16 Jan 2005, Andrew Hutchings wrote: > > >>Is there any way to rediect the request http://www.domain.com to >>$DOCROOT/domain.com/www >>Or is it something I should write a module / patch for? >> >>Regards >>Andrew >> > > -- Andrew Hutchings Systems Operator / Developer / NWLUG Admin (http://www.nwlug.org.uk) Netserve Consultants Ltd. http://www.domaincity.co.uk/ From shiner_chen at yahoo.com.cn Sat Jan 22 07:44:07 2005 From: shiner_chen at yahoo.com.cn (shiner chen) Date: Sat, 22 Jan 2005 15:44:07 +0800 (CST) Subject: how does tux process dyn-static hybrid request Message-ID: <20050122074407.15020.qmail@web15502.mail.cnb.yahoo.com> I know that tux will redirect the clinet's dynamic rquest to secondary web server, i.e apache. But if client just send another static request atfer dyn. one whithin keep-alive, then that static req. must be processed by apache, not tux. As a result, tux's specweb99 test result maybe similar with apache. That is my actual specweb99 test result. My question is how tux could process dyn-static hybrid test? can tux process static req. after dyn. one continously with one keep-alive socket? _________________________________________________________ Do You Yahoo!? ??????????????? http://cn.rd.yahoo.com/mail_cn/tag/1g/*http://cn.mail.yahoo.com/ From da_monumental_1 at nospammail.net Sat Jan 22 09:18:06 2005 From: da_monumental_1 at nospammail.net (Jarrett Graham) Date: Sat, 22 Jan 2005 04:18:06 -0500 Subject: Can't enable debug Message-ID: <1106385486.1531.213319415@webmail.messagingengine.com> I trying to debug tux to see why it won't execute my cgi scripts. I've followed various instructions on the list to get it working. Using a vanilla 2.6.10 kernel, I applied the tux3-2.6.10-rc3-A7 patch. When compiling the kernel, I get this error. CC net/tux/accept.o CC net/tux/input.o CC net/tux/userspace.o CC net/tux/cachemiss.o CC net/tux/output.o CC net/tux/redirect.o CC net/tux/postpone.o CC net/tux/logger.o CC net/tux/proto_http.o CC net/tux/proto_ftp.o CC net/tux/proc.o CC net/tux/main.o CC net/tux/mod.o CC net/tux/abuf.o net/tux/main.c: In function `__sys_tux': net/tux/main.c:1049: error: structure has no member named `rlim' make[2]: *** [net/tux/main.o] Error 1 make[1]: *** [net/tux] Error 2 Is there a fix for this? Also, I assume that cgi.c is the part of tux that processes cgi scripts. However, it is only compiled when "External CGI" is enabled in the kernel config. Is External CGI meant for tux interacting with custom build tux modules? My config: tux.mime.types: TUX/CGI cgi pl pm TUX/module tux x TUX/redirect redir application/andrew-inset ez application/mac-binhex40 hqx application/mac-compactpro cpt application/msword doc application/octet-stream bin dms lha lzh exe class application/oda oda application/pdf pdf application/postscript ai eps ps application/rtf rtf application/smil smi smil application/x-bcpio bcpio application/x-bzip2 bz2 application/x-cdlink vcd application/x-chess-pgn pgn application/x-cpio cpio application/x-csh csh application/x-director dcr dir dxr application/x-dvi dvi application/x-futuresplash spl application/x-gtar gtar application/x-gzip gz tgz application/x-hdf hdf application/x-javascript js application/x-koan skp skd skt skm application/x-latex latex application/x-netcdf nc cdf application/x-rpm rpm application/x-sh sh application/x-shar shar application/x-shockwave-flash swf application/x-stuffit sit application/x-sv4cpio sv4cpio application/x-sv4crc sv4crc application/x-tar tar application/x-tcl tcl application/x-tex tex application/x-texinfo texinfo texi application/x-troff t tr roff application/x-troff-man man application/x-troff-me me application/x-troff-ms ms application/x-ustar ustar application/x-wais-source src application/zip zip audio/basic au snd audio/midi mid midi kar audio/mpeg mpga mp2 mp3 audio/x-aiff aif aiff aifc audio/x-pn-realaudio ram rm audio/x-realaudio ra audio/x-wav wav chemical/x-pdb pdb xyz image/gif gif image/ief ief image/jpeg jpeg jpg jpe image/png png image/tiff tiff tif image/x-cmu-raster ras image/x-portable-anymap pnm image/x-portable-bitmap pbm image/x-portable-graymap pgm image/x-portable-pixmap ppm image/x-rgb rgb image/x-xbitmap xbm image/x-xpixmap xpm image/x-xwindowdump xwd model/iges igs iges model/mesh msh mesh silo model/vrml wrl vrml text/css css text/plain asc txt text/richtext rtx text/rtf rtf text/sgml sgml sgm text/tab-separated-values tsv text/x-setext etx text/xml xml video/mpeg mpeg mpg mpe video/quicktime qt mov video/x-msvideo avi video/x-sgi-movie movie x-conference/x-cooltalk ice text/html html htm application/xhtml+xml xhtml xht /etc/conf.d/tux: # Config file for /etc/init.d/tux # modified from original to be Gentoo like # $Header: /var/cvsroot/gentoo-x86/www-servers/tux/files/tux.conf.d,v 1.1 2004/08/08 19:03:31 stuart Exp $ # TUX_THREADS sets the number of kernel threads (and associated daemon # threads) that will be used. $TUX_THREADS defaults to 1. # TUX_THREADS=1 # DOCROOT is the document root; it works the same way as other web # servers such as apache. This must have only 1 trailing /. TUX_DOCROOT=/var/www/ # LOGFILE is the file where tux logs information for each # request. Note that tux writes log files in a binary format and to # read them you will need to first convert them into standard # W3C-conforming HTTPD log files using the utility tux2w3c. If no # LOGFILE is specified then requests will not be logged. TUX_LOGFILE=/var/log/tux # TUX_UID and TUX_GID are the user and group as which the daemon runs # This does not mean that you should execute untrusted modules -- they # are opened as user/group root, which means that the _init() function, # if it exists, is run as root. This feature is only designed to help # protect from programming mistakes; it is NOT really a security mechanism. TUX_UID=81 TUX_GID=81 # CGIs can be started in a chroot environment by default. # set TUX_CGIROOT=/ if you want CGI programs to have access to the whole system. #TUX_CGIROOT=/home/httpd/htdocs TUX_CGIROOT=/ # each HTTP connection has an individual timer that makes sure # no connection hangs forever. (due to browser bugs or DoS attacks.) # TUX_KEEPALIVE=30 # TUX_MODULES is a list of user-space TUX modules. User-space TUX # modules are used to serve dynamically-generated data via tux. # "man 2 tux" for more information # TUX_MODULES="demo.tux demo2.tux demo3.tux demo4.tux" # TUX_MODULEPATH is the path to user-space TUXapi modules # TUX_MODULEPATH="/" CGI_UID=81 CGI_GID=81 /etc/sysctl.conf: # /etc/sysctl.conf: # $Header: /home/cvsroot/gentoo-src/rc-scripts/etc/sysctl.conf,v 1.3 2002/11/18 19:39:22 azarah Exp $ # Disables packet forwarding #net.ipv4.ip_forward = 0 # Disables IP dynaddr #net.ipv4.ip_dynaddr = 0 # Disable ECN net.ipv4.tcp_ecn = 0 # Enables source route verification #net.ipv4.conf.default.rp_filter = 1 # Disables the magic-sysrq key #kernel.sysrq = 0 #fs.binfmt_misc.register = :WINEXE:M::MZ::/usr/bin/cedega: fs.file-max = 65535 vm.swappiness = 40 dev.rtc.max-user-freq = 1024 #Tux Webserver #net.tux.compression = 1 net.tux.virtual_server = 1 #net.tux.max_connect = 10000 -- Jarrett Graham da_monumental_1 at nospammail.net From sitz at onastick.net Sat Jan 22 18:46:20 2005 From: sitz at onastick.net (Noah) Date: Sat, 22 Jan 2005 13:46:20 -0500 Subject: how does tux process dyn-static hybrid request In-Reply-To: <20050122074407.15020.qmail@web15502.mail.cnb.yahoo.com> References: <20050122074407.15020.qmail@web15502.mail.cnb.yahoo.com> Message-ID: <20050122184620.GA24675@radu.onastick.net> On Sat, Jan 22, 2005 at 03:44:07PM +0800, shiner chen wrote: > My question is how tux could process dyn-static hybrid > test? can tux process static req. after dyn. one > continously with one keep-alive socket? I don't think you can. What you /can/ do (IIRC) is disable HTTP keep-alives in apache (via the KeepAlive directive); the downside to this is that apache closes the connection after each request. The upside is that tux has a shot at every request. It's up to the admin to determine whether this is worth it, and that will depend on the traffic profile of the site(s) being served. --n -- dd of=/dev/fd0 if=/dev/flippy bs=1024 ^^^ Making Flippy Floppy From da_monumental_1 at nospammail.net Mon Jan 24 08:48:27 2005 From: da_monumental_1 at nospammail.net (Jarrett Graham) Date: Mon, 24 Jan 2005 03:48:27 -0500 Subject: Kernel Oops Message-ID: <1106556507.11655.213414212@webmail.messagingengine.com> I got this oops. kernel BUG at fs/namei.c:498! invalid operand: 0000 [#1] PREEMPT Modules linked in: snd_seq_midi snd_pcm_oss snd_mixer_oss snd_seq_oss snd_seq_midi_event snd_seq snd_ens1371 snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd snd_page_alloc tuner msp3400 i2c_core i810 intel_agp agpgart e100 mii sbp2 ohci1394 ieee1394 ohci_hcd uhci_hcd usb_storage scsi_mod usbhid ehci_hcd usbcore CPU: 0 EIP: 0060:[] Not tainted VLI EFLAGS: 00210292 (2.6.10-gentoo-r6) EIP is at link_path_walk+0xa19/0xd70 eax: c3df5a00 ebx: dbf37dfc ecx: c5db6af4 edx: 00000000 esi: dbf36000 edi: c367c62c ebp: dbf37e40 esp: dbf37db4 ds: 007b es: 007b ss: 0068 Process tux (pid: 21572, threadinfo=dbf36000 task=c3df5a00) Stack: dbf37dfc dbf37e00 dbf37dfc 00000040 0000023f c5db6af4 0000023f ffffffd8 dfeb7760 ce235444 dbf36000 00000001 00000000 0000023f dbf37dfc 0000023f 00000001 00000040 dfeb4e00 c5db6af4 329fa116 00000007 dd47b1b0 c03c840b Call Trace: [] read_request+0x11b/0x170 [] __tux_lookup+0x10/0x30 [] tux_lookup+0xb3/0x110 [] lookup_url+0x70/0x5f0 [] wake_up_process+0x1d/0x30 [] http_lookup_vhost+0xa0/0x130 [] http_process_message+0x6c/0x340 [] tux_schedule_atom+0x1d/0x30 [] process_requests+0xc7/0xf0 [] event_loop+0x138/0x1a0 [] __sys_tux+0x1b2/0x880 [] sys_chdir+0x76/0x90 [] sys_setuid+0xab/0xd0 [] syscall_call+0x7/0xb Code: ff 28 00 eb 83 0f 0b 19 01 22 69 40 c0 e9 60 ff ff ff 89 2c 24 e8 28 f1 ff ff 89 f8 e9 0d fe ff ff e8 1c 03 29 00 e9 68 fd ff ff <0f> 0b f2 01 94 9f 40 c0 e9 50 fd ff ff 89 2c 24 e8 02 f1 ff ff <5>TUX: thread 0 stopping ... -- Jarrett Graham da_monumental_1 at nospammail.net From bbsea at mac.com Tue Jan 25 02:46:17 2005 From: bbsea at mac.com (bbsea at mac.com) Date: Mon, 24 Jan 2005 18:46:17 -0800 Subject: tux load/performance questions Message-ID: <9232212.1106621177288.JavaMail.bbsea@mac.com> Hello, I have recently started working with TUX and was forced to use it in a production environment before I had the chance to become completely familiar with it. Here is the scenario.. This webserver is dedicated to serving jpg images only. Approx 500,000 images ranging in size from 1k to 40k single cpu (Intel(R) Pentium(R) 4 CPU 2.40GHz) 2 GB RAM 3ware hardware IDE RAID 0 80GB Linux 2.4.21-9.EL (Redhat Enterprise ES) tux-3.2.14-1 issue: server can not currently push more than 4 mbps (as measured by iptraf) I am reluctant to throw hardware at this problem until I can determine that it is necessary. I would of not expected to see this sort of load while pushing only 4 mbps of traffic. Since this is a production environment I also am hesitant about 'trying' different software versions on the fly.. thoughts? diagnostic output taken under peak production load: more information available upon request of course.. 18:23:36 up 1 day, 5:41, 2 users, load average: 7.96, 5.15, 3.93 Mem: 2063012k av, 2041188k used, 21824k free nobody 15 0 756 756 644 S 18.9 0.0 111:11 0 async IO 0/2 nobody 16 0 756 756 644 R 14.9 0.0 205:05 0 async IO 0/1 nobody 15 0 756 756 644 S 10.9 0.0 47:08 0 async IO 0/3 nobody 15 0 756 756 644 S 6.3 0.0 18:17 0 async IO 0/4 nobody 15 0 756 756 644 S 0.9 0.0 0:38 0 async IO 0/8 nobody 15 0 756 756 644 S 0.9 0.0 0:07 0 async IO 0/11 nobody 15 0 756 756 644 S 0.7 0.0 10:48 0 tux Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn sda 77.49 69.59 784.97 7452148 84059990 sda1 0.00 0.00 0.00 122 46 sda2 0.00 0.01 0.02 1440 2128 sda3 77.48 69.57 784.95 7449642 84057816 sdb 0.00 0.01 0.00 568 0 sdb1 0.00 0.00 0.00 88 0 procs memory swap io system cpu r b swpd free buff cache si so bi bo in cs us sy id wa 1 0 4 15252 138180 1695272 0 0 35 393 238 365 0 26 63 10 Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s rkB/s wkB/s avgrq-sz avgqu-sz await svctm %util sda 0.01 29.33 8.69 68.81 69.58 785.17 34.79 392.58 11.03 0.33 4.21 1.68 13.02