general/3386: Memory usage per httpd-process rockets up to 100MB
Number: 3386 Category: general Synopsis: Memory usage per httpd-process rockets up to 100MB Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Thu Nov 12 06:40:00 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3 (1030399) Environment: Sun Solaris 2.6 mySQL 3.21.33 PHP 3.0.5 Apache API Version: 19980917 Description: Max Requests: per child: 30keep alive: onmax per connection: 100 Timeouts: connection: 300keep-alive: 15 Loaded modules: mod_auth, mod_php3, mod_auth_mysql, mod_auth_radius, mod_setenvif, mod_headers, mod_expires, mod_access, mod_alias, mod_userdir, mod_actions, mod_imap, mod_asis, mod_cgi, mod_dir, mod_autoindex, mod_include, mod_status, mod_negotiation, mod_mime, mod_log_referer, mod_log_agent, mod_log_config, mod_env, http_core With this configuration we experience serious perfomance problems. Each process seems to use more and more memory (peak value today: 480MB) until all of the system memory (512 MB) is used and the system starts to swap continously, ruining the overall perfomance. Could this be a mySQL-related problem, e.g. infinite loops or something like that ?! We searched several mailing lists but no-one seems to have this problem, as nearly all memory leak- problems are related to the rewrite-mod, which we are appearantly not using. If there is someone having a solution, he should mail us as fast as possible. How-To-Repeat: Sorry, but encountering our poor perfomance would surly not help. Fix: None at all. Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and restart httpd.
Number: 3387 Category: os-linux Synopsis: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and restart httpd. Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Thu Nov 12 08:10:01 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.1 Environment: Linux eddie.sccd.ctc.edu 2.0.34 #3 Wed Aug 26 15:23:44 PDT 1998 i686 unknown redhat 5.1, apache 1.3.1, gcc-2.7.2.3-11, glibc-debug-2.0.7-19, glibc-profile-2.0.7-19, glibc-devel-2.0.7-19, glibc-2.0.7-19 Description: I cannot pinpoint any specific time or reason that the server dies. Just every 6 or 7 hours the all the the httpd children will slowly die (within about 10 minutes) until there is only one process left, the one owned by root. I then stop and restart the server using the method described on apaches web page ( kill -TERM `cat /var/run/httpd.pid`). About every three days this method of stopping and restarting the server does not work because it says the port (80) is still in use and there is no /var/run/httpd.pid file and I must reboot the server. I have recompiled httpd recently and it only caused the problem to be worse. I also increased min and max spare servers so as to ensure the server was not getting overloaded. But this problem does not seem to be related to the nubmer of processes running on the system. I'm also attaching a process and netstat report from when the server dies. Don't know if it will help or not, but I hope so. - server up, 11 processes, at 18:48:00Monday server up, 11 processes, at 18:50:01Monday server up, 11 processes, at 18:52:00Monday server up, 10 processes, at 18:54:01Monday server up, 10 processes, at 18:56:00Monday server up, 10 processes, at 18:58:01Monday server up, 6 processes, at 19:00:00Monday PS output root 1160 0.0 0.2 848 348 ? S19:00 0:00 grep httpd root 31731 0.0 0.6 1648 880 ? S16:21 0:00 httpd wuser 782 0.0 0.8 1764 1052 ? S18:32 0:00 httpd wuser 785 0.0 0.8 1764 1052 ? S18:32 0:00 httpd wuser 789 0.0 0.8 1764 1052 ? S18:33 0:00 httpd wuser 790 0.0 0.8 1764 1052 ? S18:33 0:00 httpd wuser 791 0.0 0.8 1764 1052 ? S18:33 0:00 httpd wuser 792 0.0 0.8 1764 1052 ? S18:33 0:00 httpd Netstat output tcp0 0 *:www *:* LISTEN tcp0 0 eddie.sccd.ctc.edu:www cs207-21.student.w:1034 FIN_WAIT2 tcp0 9003 eddie.sccd.ctc.edu:www cs207-21.student.w:1035 FIN_WAIT1 tcp0 3327 eddie.sccd.ctc.edu:www cs207-21.student.w:1036 FIN_WAIT1 tcp1 0 eddie.sccd.ctc.edu:www 134.39.67.90:1888 TIME_WAIT tcp1 0 eddie.sccd.ctc.edu:www 134.39.67.90:1889 TIME_WAIT tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.240:1438 ESTABLISHED tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.240:1439 ESTABLISHED tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.240:1440 ESTABLISHED tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.240:1441 ESTABLISHED server up, 1 processes, at 19:01:00Monday PS output root 1177 0.0 0.2 848 348 ? S19:01 0:00 grep httpd root 31731 0.0 0.6 1648 880 ? S16:21 0:00 httpd wuser 792 0.0 0.8 1764 1052 ? S18:33 0:00 httpd Netstat output tcp0 0 *:www *:* LISTEN tcp0 0 eddie.sccd.ctc.edu:www cs207-21.student.w:1034 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www cs207-21.student.w:1035 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www cs207-21.student.w:1036 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.240:1438 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.240:1440 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.240:1441 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.80:1439 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.80:1441 FIN_WAIT2 server up, 1 processes, at 19:02:01Monday PS output root 1194 0.0 0.4 1160 624 ? R19:02 0:00 grep httpd root 31731 0.0 0.6 1648 880 ? S16:21 0:00 httpd wuser 792 0.0 0.8 3248 1096 ? S18:33 0:00 httpd Netstat output tcp0 0 *:www *:* LISTEN tcp0 0 eddie.sccd.ctc.edu:www cs207-21.student.w:1035 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www cs207-21.student.w:1036 FIN_WAIT2 tcp0 0 eddie.sccd.ctc.edu:www 134.39.67.145:1451 FIN_WAIT2 tcp0 38652 eddie.sccd.ctc.edu:www 134.39.67.145:1452 ESTABLISHED server
os-windows/3388: Bluescreen with VxD error message similar to PR#3220. It happens when my server is in peak hours.
Number: 3388 Category: os-windows Synopsis: Bluescreen with VxD error message similar to PR#3220. It happens when my server is in peak hours. Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: support Submitter-Id: apache Arrival-Date: Thu Nov 12 08:20:01 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3 Environment: Win95b. Use the binary code directly. Description: I have a very high traffic to this PC, more than 50 hits in peak hours. I have taken the advice and upgrade to winsocket2. And to avoid the FIN_WAIT_2 problem, I set the keep-alive to Off. This two measures seem to improve the performance a little bit, since the server survived the peak hours last time. But yesterday it's down with blue screen. I checked the stat report of my website. It shows that the server was dead at peak hours. I use the Win95 Apache for http server only. And I have tried using other windows webservers. I met the same problem. So here is a question. How can I improve the stability of the system when using Apache. What other parameters should I change to control the traffic? I've viewd the PR database carefully. And 3220 is the only related one. I also know that it's still open. Hope to get some solutions soon. How-To-Repeat: I think if you have Win95b, winsock1, and a very high traffic, you would be very easy to get the bluescreen in peak hours. Fix: Don't know. Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: mod_digest/3385: digest authentication return 401 413 where it should be ok
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: digest authentication return 401 413 where it should be ok State-Changed-From-To: open-closed State-Changed-By: marc State-Changed-When: Thu Nov 12 09:26:54 PST 1998 State-Changed-Why: The digest auth that Apache supports is the digest auth, as defined a few years ago. It will not work with IE's implementation, since IE's implementation is of a different standard. At some point, Apache's mod_digest should be upgraded to that new standard.
Re: pending/3384: Perl 5.005_02 will not run under Apache 1.3.3 win 95
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Perl 5.005_02 will not run under Apache 1.3.3 win 95 State-Changed-From-To: open-closed State-Changed-By: marc State-Changed-When: Thu Nov 12 09:28:34 PST 1998 State-Changed-Why: Closing second misfiled PR on this subject; already dealt with in the original PR. Responsible-Changed-From-To: gnats-admin-apache Responsible-Changed-By: marc Responsible-Changed-When: Thu Nov 12 09:28:34 PST 1998 Responsible-Changed-Why: .
RE: os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and r estart httpd.
The following reply was made to PR os-linux/3387; it has been noted by GNATS. From: Marxen, Sara [EMAIL PROTECTED] To: '[EMAIL PROTECTED]' [EMAIL PROTECTED], '[EMAIL PROTECTED]' [EMAIL PROTECTED] Cc: Subject: RE: os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and r estart httpd. Date: Thu, 12 Nov 1998 09:21:27 -0800 There is another thing I meant to report. My Minspareservers is set to 10. But the number of server processes regularly goes below 10, for example last time before it died. the number of httpd processes waiting was below 10 for over 35 minutes. Other times the server will go down to about 4 or 6 servers and stay there for a half hour before returning to the minspare server amount of 10. Isn't the min spare servers suppossed to signify the least amount of spare servers that should be running? So there should be no less than 10 for more than a few minutes? -- From:[EMAIL PROTECTED]:[EMAIL PROTECTED] Reply To:[EMAIL PROTECTED];[EMAIL PROTECTED] Sent:Thursday, November 12, 1998 10:10 AM To: [EMAIL PROTECTED] Subject: Re: os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and restart httpd. Thank you very much for your problem report. It has the internal identification `os-linux/3387'. The individual assigned to look at your report is: apache. Category: os-linux Responsible:apache Synopsis: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and restart httpd. Arrival-Date: Thu Nov 12 08:10:01 PST 1998
Re: os-windows/3388: Bluescreen with VxD error message similar to PR#3220. It happens when my server is in peak hours.
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Bluescreen with VxD error message similar to PR#3220. It happens when my server is in peak hours. State-Changed-From-To: open-closed State-Changed-By: marc State-Changed-When: Thu Nov 12 09:30:48 PST 1998 State-Changed-Why: I'm afraid we really can't help you much. You say you have the same problem with other web servers, indicating that it is likely a problem with 95. 95 is not designed to be a server and will not operate well under load. I am not suprised that Win95 can't handle a significant load.
Re: config/3281: Unable to get .htaccess to work at subdirec (fwd)
The following reply was made to PR config/3281; it has been noted by GNATS. From: Marc Slemko [EMAIL PROTECTED] To: Apache bugs database [EMAIL PROTECTED] Cc: Subject: Re: config/3281: Unable to get .htaccess to work at subdirec (fwd) Date: Thu, 12 Nov 1998 09:30:53 -0800 (PST) -- Forwarded message -- Date: Thu, 12 Nov 1998 12:56:37 + From: Mark Cousins [EMAIL PROTECTED] To: Marc Slemko [EMAIL PROTECTED] Subject: Re: config/3281: Unable to get .htaccess to work at subdirec Marc, In reply to your prior message there are some areas I do not fully understand so in order to help you to help us here is a copy of our access.conf file to suggest any possible solutions. In reply to your questions ? Is this in a Directory container? I am not sure what a directory container is ? If so, what what? Where on the filesystem is the htaccess file located? All files as per access.conf for web are in /home/httpd/html/ so have created a suddir under html called markc and put files in there. .htaccess file is located in /home/httpd/ = does work here. ( prevents access to main page ) and located in /home/httpd/html/markc/ = does not work here. Are you sure the Directory container in question covers that location? As I am not sure about directory container I do not know how this directory can be covered by .htaccess file and rules. Or where to make thwe relevent entry in access.conf. === === ACCESS.CONF # This should be changed to whatever you set DocumentRoot to. Directory /home/httpd/html # This may also be None, All, or any combination of Indexes, # Includes, FollowSymLinks, ExecCGI, or MultiViews. # Note that MultiViews must be named *explicitly* --- Options All # doesn't give it to you (or at least, not yet). Options Indexes FollowSymLinks # This controls which options the .htaccess files in directories can # override. Can also be All, or any combination of Options, FileInfo, # AuthConfig, and Limit #AllowOverride None AllowOverride AuthConfig # Controls who can get stuff from this server. order allow,deny allow from all /Directory # /usr/local/etc/httpd/cgi-bin should be changed to whatever your ScriptAliased # CGI directory exists, if you have that configured. # Directory /home/httpd/cgi-bin Directory /home/httpd/html AllowOverride none Options All /Directory # Allow server status reports, with the URL of http://servername/server-status # Change the .your_domain.com to match your domain to enable. #Location /server-status #SetHandler server-status #order deny,allow #deny from all #allow from .your_domain.com #/Location # There have been reports of people trying to abuse an old bug from pre-1.1 # days. This bug involved a CGI script distributed as a part of Apache. # By uncommenting these lines you can redirect these attacks to a logging # script on phf.apache.org. Or, you can record them yourself, using the script # support/phf_abuse_log.cgi. Location /cgi-bin/phf* deny from all ErrorDocument 403 http://phf.apache.org/phf_abuse_log.cgi /Location # You may place any other directories or locations you wish to have # access information for after this one. Cheers Mark.Cousins ---[Mail message history]--- To: Mark Cousins/Northern Europe/London/SITA/WW cc: Apache bugs database [EMAIL PROTECTED] From: Marc Slemko [EMAIL PROTECTED] Date: 29/10/98 16:44:53 Subject: Re: config/3281: Unable to get .htaccess to work at subdirec On Thu, 29 Oct 1998, Mark Cousins wrote: Ryan, have checked the access.conf file and the entry is as follows: # This controls which option .htaccess files in directories can override. # can also be All, any combination of options, fileinfo, authconfig, and limit #allowoveride none allowoveride authconfig Is this in a Directory container? If so, what what? Where on the filesystem is the htaccess file located? Are you sure the Directory container in question covers that location? --
Re: config/3281: Unable to get .htaccess to work at subdirec
The following reply was made to PR config/3281; it has been noted by GNATS. From: Marc Slemko [EMAIL PROTECTED] To: Mark Cousins [EMAIL PROTECTED] Cc: Apache bugs database [EMAIL PROTECTED] Subject: Re: config/3281: Unable to get .htaccess to work at subdirec Date: Thu, 12 Nov 1998 09:32:36 -0800 (PST) On Thu, 12 Nov 1998, Mark Cousins wrote: Marc, In reply to your prior message there are some areas I do not fully understand so in order to help you to help us here is a copy of our access.conf file to suggest any possible solutions. In reply to your questions ? Is this in a Directory container? I am not sure what a directory container is ? A Directory container is a Directory foo... then a few lines then /Directory bit in the config file. It makes restrictions on a per directory basis. Directory /home/httpd/html # This may also be None, All, or any combination of Indexes, # Includes, FollowSymLinks, ExecCGI, or MultiViews. # Note that MultiViews must be named *explicitly* --- Options All # doesn't give it to you (or at least, not yet). Options Indexes FollowSymLinks # This controls which options the .htaccess files in directories can # override. Can also be All, or any combination of Options, FileInfo, # AuthConfig, and Limit #AllowOverride None AllowOverride AuthConfig Go through the docs on the AllowOverride directive. Here, it says that AuthConfig is allowed, which is what you appear to want to do in your htaccess file. [...] Directory /home/httpd/html AllowOverride none Options All /Directory Then this section, for the html directory, says that no overrides are allowed which means that htaccess files will be ignored in that directory, which is the behaviour you are seeing.
Re: os-windows/3359: Configuring Apache only for proxy cache it report a runtime error.
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: Configuring Apache only for proxy cache it report a runtime error. State-Changed-From-To: open-closed State-Changed-By: marc State-Changed-When: Thu Nov 12 09:42:01 PST 1998 State-Changed-Why: User config error.
other/3391: REMOTE_HOST environment variable poorly documented and treated
Number: 3391 Category: other Synopsis: REMOTE_HOST environment variable poorly documented and treated Confidential: no Severity: serious Priority: medium Responsible:apache State: open Class: change-request Submitter-Id: apache Arrival-Date: Thu Nov 12 09:50:00 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3 Environment: AIX chywolf 2 4 000124A24C00 Description: The REMOTE_HOST variable passed to CGIs on NCSA used to have the remote host of a connection, or, if no host was available, the IP address. Newer versions of Apache has changed that to: REMOTE_HOST exists only if the- re *is* a host to be informed. However, there's still no note about that in the compatibility notes. I may be showing myself selfish, but a lot of CGIs created for NCSA uses the REMOTE_HOST variable and they fail when it's not set, as describbed in pre- vious bug report(s). How-To-Repeat: Fix: I'd like to suggest that the REMOTE_HOST would still acquire the IP address if there's no host name available, but the REMOTE_ADDR would still exist. It would only be a matter of checking if REMOTE_HOST == REMOTE_ADDR to know if there *is* a host name or not... and Apache would retain a somewhat nice com- patibility with the old nice NCSA. Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: other/3391: REMOTE_HOST environment variable poorly documented and treated
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: REMOTE_HOST environment variable poorly documented and treated State-Changed-From-To: open-closed State-Changed-By: marc State-Changed-When: Thu Nov 12 09:52:17 PST 1998 State-Changed-Why: The CGI spec (essentially from NCSA) says that the REMOTE_HOST variable should not be set unless there is a reverse lookup. This is described quite clearly in the CHANGES file and in the upgrading_to_1.3 file.
RE: os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and r estart httpd.
I also thought the following would be helpful: This is the error log from this morning, it is confusing to me that the information messages say, for example, spawning 8, 9 idle, 49 total children while the ps ouput at that time shows only 18 processes. Attached are the error log for a certain period and below that is the number of proccesses running at each time corresponding to the error log. ERROR LOG OUTPUT [Thu Nov 12 08:18:18 1998] [info] server seems busy, (you may need to increase StartServers, or Min/MaxSpa reServers), spawning 8 children, there are 9 idle, and 49 total children [Thu Nov 12 08:44:06 1998] [crit] (98)Address already in use: make_sock: could not bind to port 80 [Thu Nov 12 08:51:17 1998] [info] server seems busy, (you may need to increase StartServers, or Min/MaxSpa reServers), spawning 8 children, there are 7 idle, and 64 total children [Thu Nov 12 09:43:10 1998] [crit] (98)Address already in use: make_sock: could not bind to port 80 [Thu Nov 12 09:44:07 1998] [crit] (98)Address already in use: make_sock: could not bind to port 80 [Thu Nov 12 09:44:54 1998] [notice] httpd: caught SIGTERM, shutting down [Thu Nov 12 09:45:00 1998] [info] created shared memory segment #768 [Thu Nov 12 09:45:00 1998] [notice] Apache/1.3.1 (Unix) configured -- resuming normal operations [Thu Nov 12 09:45:00 1998] [info] Server built: Nov 5 1998 14:56:18 NUMBER OF PROCESSE RUNNING AT THE SAME TIME server up, 18 processes, at 08:18:00Thursday server up, 32 processes, at 08:44:00Thursday server up, 32 processes, at 08:51:01Thursday server up, 1 processes, at 09:43:01Thursday server up, 1 processes, at 09:44:01Thursday server up, 20 processes, at 09:45:02Thursday NOTE: Ther server stopped actually serving at around 9:40 though there were processes still running. Also the number of processes in the list above does not include in the count the parent process which is owned by root. Is there any other info that would be useful? An strace? -- From: [EMAIL PROTECTED]:[EMAIL PROTECTED] Reply To: [EMAIL PROTECTED];[EMAIL PROTECTED] Sent: Thursday, November 12, 1998 10:10 AM To: [EMAIL PROTECTED] Subject: Re: os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and restart httpd. Thank you very much for your problem report. It has the internal identification `os-linux/3387'. The individual assigned to look at your report is: apache. Category: os-linux Responsible:apache Synopsis: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and restart httpd. Arrival-Date: Thu Nov 12 08:10:01 PST 1998
RE: os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and r estart httpd.
The following reply was made to PR os-linux/3387; it has been noted by GNATS. From: Marxen, Sara [EMAIL PROTECTED] To: '[EMAIL PROTECTED]' [EMAIL PROTECTED], '[EMAIL PROTECTED]' [EMAIL PROTECTED] Cc: Subject: RE: os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and r estart httpd. Date: Thu, 12 Nov 1998 10:14:54 -0800 I also thought the following would be helpful: This is the error log from this morning, it is confusing to me that the information messages say, for example, spawning 8, 9 idle, 49 total children while the ps ouput at that time shows only 18 processes. Attached are the error log for a certain period and below that is the number of proccesses running at each time corresponding to the error log. ERROR LOG OUTPUT [Thu Nov 12 08:18:18 1998] [info] server seems busy, (you may need to increase StartServers, or Min/MaxSpa reServers), spawning 8 children, there are 9 idle, and 49 total children [Thu Nov 12 08:44:06 1998] [crit] (98)Address already in use: make_sock: could not bind to port 80 [Thu Nov 12 08:51:17 1998] [info] server seems busy, (you may need to increase StartServers, or Min/MaxSpa reServers), spawning 8 children, there are 7 idle, and 64 total children [Thu Nov 12 09:43:10 1998] [crit] (98)Address already in use: make_sock: could not bind to port 80 [Thu Nov 12 09:44:07 1998] [crit] (98)Address already in use: make_sock: could not bind to port 80 [Thu Nov 12 09:44:54 1998] [notice] httpd: caught SIGTERM, shutting down [Thu Nov 12 09:45:00 1998] [info] created shared memory segment #768 [Thu Nov 12 09:45:00 1998] [notice] Apache/1.3.1 (Unix) configured -- resuming normal operations [Thu Nov 12 09:45:00 1998] [info] Server built: Nov 5 1998 14:56:18 NUMBER OF PROCESSE RUNNING AT THE SAME TIME server up, 18 processes, at 08:18:00Thursday server up, 32 processes, at 08:44:00Thursday server up, 32 processes, at 08:51:01Thursday server up, 1 processes, at 09:43:01Thursday server up, 1 processes, at 09:44:01Thursday server up, 20 processes, at 09:45:02Thursday NOTE: Ther server stopped actually serving at around 9:40 though there were processes still running. Also the number of processes in the list above does not include in the count the parent process which is owned by root. Is there any other info that would be useful? An strace? -- From:[EMAIL PROTECTED]:[EMAIL PROTECTED] Reply To:[EMAIL PROTECTED];[EMAIL PROTECTED] Sent:Thursday, November 12, 1998 10:10 AM To: [EMAIL PROTECTED] Subject: Re: os-linux/3387: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and restart httpd. Thank you very much for your problem report. It has the internal identification `os-linux/3387'. The individual assigned to look at your report is: apache. Category: os-linux Responsible:apache Synopsis: Regularly the httpd children will all die with no errors given, the root process will still be running, must stop and restart httpd. Arrival-Date: Thu Nov 12 08:10:01 PST 1998
mod_cgi/3392: putting leading spaces after ? in Query string places QS values in other ENV variables
Number: 3392 Category: mod_cgi Synopsis: putting leading spaces after ? in Query string places QS values in other ENV variables Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Thu Nov 12 11:50:01 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.1 Environment: BSDI 3.1 Description: NOTE, I am not reporting a problem with a non-supported cgi-script. There seems to be a problem with embedding leading spaces after the initial ? or after any in QUERY_STRING input. If you still have access to the test-cgi script from the cgi-pack or if you hae a simple cgi environment variable displayer do the following visit the script http://www.yourplace.com/test-cgi it works fine. Append to it ?a=b or ?a=bc=d it will work fas expected. If you instead append, ?a=b or ?a=bc=d, that part that begins with the whitespace is shoved into the SERVER_PROTOCOL variable [on my system at least). I have not had time to look at mod_cgi or whatever other modules might handle the splitting up of the environment but this would not seem to be correct behaviour How-To-Repeat: See above Fix: not at this time Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
Re: mod_cgi/3392: putting leading spaces after ? in Query string places QS values in other ENV variables
[In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ] Synopsis: putting leading spaces after ? in Query string places QS values in other ENV variables State-Changed-From-To: open-closed State-Changed-By: marc State-Changed-When: Thu Nov 12 11:51:22 PST 1998 State-Changed-Why: Unencoded spaces are not valid in a request. All spaces must be % encoded (to %20). Spaces are used to seperate the method, the request URI, and the protocol. Apache's behaviour is correct.
documentation/3394: undocumented command in dbmmanage.1
Number: 3394 Category: documentation Synopsis: undocumented command in dbmmanage.1 Confidential: no Severity: non-critical Priority: medium Responsible:apache State: open Class: doc-bug Submitter-Id: apache Arrival-Date: Thu Nov 12 13:50:01 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3 Environment: any Description: The add command is not documented, but is important, especially for use from scripts. The perl4 version of dbmmanage supported reading the encrypted password from the command line, so people need to know how to do that using the new version. How-To-Repeat: man dbmmanage Fix: --- dbmmanage.1 1998/10/19 08:53:46 1.1.1.1 +++ dbmmanage.1 1998/11/12 21:37:35 @@ -60,6 +60,8 @@ .I command ] [ .I username +] [ +.I encpassword ] .PP .SH DESCRIPTION @@ -87,6 +89,10 @@ The filename of the DBM format file. Usually without the extension .pag or .dir. .IP \fB\fIcommand\fP This selects the operation to perform: +.TP 12 +.B add +Adds an entry for \fIusername\fP to \fIfilename\fP using the encrypted +password \fIencpassword\fP. .TP 12 .B adduser Asks for a password and then adds an entry for \fIusername\fP to Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
general/3395: Problem with ProxyPass and authentification
Number: 3395 Category: general Synopsis: Problem with ProxyPass and authentification Confidential: no Severity: critical Priority: medium Responsible:apache State: open Class: sw-bug Submitter-Id: apache Arrival-Date: Thu Nov 12 14:10:00 PST 1998 Last-Modified: Originator: [EMAIL PROTECTED] Organization: apache Release:1.3.3 Environment: Unixware 2.1.2 UNIX_SV a2k 4.2MP 2.1.2 i386 x86at Description: I have two webservers. Server1 uses ProxyPass to let clients access server2. This works just fine. But when I restrict access to server1 and the client tries to access server2, the client gets 407 Proxy Authentification required What is wrong??? This worked just fine with Apache 1.2.5. I use the following configuration: ServerType standalone HostnameLookups off User nobody Group nobody ServerAdmin [EMAIL PROTECTED] ServerRoot /usr/htroot DocumentRoot /usr/htdocs ErrorLog logs/error_log CustomLog logs/access_log common ScoreBoardFile logs/apache_runtime_status ServerName web.wide.world.se ServerSignature on UseCanonicalName on KeepAlive On ProxyRequests Off ProxyPass /web2 http://web2.wide.world.se Location ~ .* Options FollowSymLinks Includes MultiViews ExecCGI AllowOverride None AuthType Basic AuthName OnLine AuthUserFile /usr/etc/passwd Limit GET POST require valid-user /Limit /Location How-To-Repeat: Set up 2 web servers. Have one proxy the other using the above configuration. Netscape reports 407 Proxy Authentification required. Fix: Audit-Trail: Unformatted: [In order for any reply to be added to the PR database, ] [you need to include [EMAIL PROTECTED] in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]
