Server :: Apache Incorrectly Handling Requests For Numeric PHP Filenames?
Nov 14, 2009
I'm working on a thorny mod_rewrite problem. I have a mac connected to my LAN running MAMP (Mac/Apache/MySQL/PHP). I request a non-existent file:
Code:
http://192.168.1.2:8888/careers/db/1.html
I see this in the mod_rewrite log file:
Code:
192.168.1.102 - - [14/Nov/2009:13:46:07 --0800] [192.168.1.2/sid#807df8][rid#8ec850/initial] (2) init rewrite engine with requested uri /careers/db/1.html
192.168.1.102 - - [14/Nov/2009:13:46:07 --0800] [192.168.1.2/sid#807df8][rid#8ec850/initial] (1) pass through /careers/db/1.html
Note that the requested uri is /careers/db/1.html
If I change just the file extension on my request to PHP like so:
Code: [URL]
Then the request uri is totally different now. Here's the rewrite log:
Code:
192.168.1.102 - - [14/Nov/2009:13:47:23 --0800] [192.168.1.2/sid#807df8][rid#8fc850/initial] (2) init rewrite engine with requested uri /Applications/MAMP/htdocs/careers/
192.168.1.102 - - [14/Nov/2009:13:47:23 --0800] [192.168.1.2/sid#807df8][rid#8fc850/initial] (1) pass through /Applications/MAMP/htdocs/careers/
Note that the requested uri now has a full path which does not include the actual filename, /Applications/MAMP/htdocs/careers/
What the heck? More info. If I request [URL], I can actually access p1.php. The requested uri is /careers/db/p1.php. The problem appears to be because the filename starts with a number. I can also request [URL] and get thru to 1.php with requested uri /careers/db//1.php. Does mod_rewrite think /1 refers to a backreference or something? Why can apache handle the html file request properly and not the php file request?
I mostly do .Net development in a Windows 7 virtualbox. I use the host for simple things such as web browsing, skype, chat, etc. All things that are fantastically available on Ubuntu which I in many ways prefer. So this has been begging the question for a while: why even use Windows on the host, seems like a Linux host would use less resources (untested) and allow my Windows VMs to run better while allowing me to do my non-development stuff in an interface I prefer.
So easiest way to do this - I downloaded wubi and installed Ubuntu. I installed in it Virtualbox, and then start add and start my VM to get this message: Failed to open a session for the virtual machine VS2010, Could not open the medium '/host/Users/George Mauer/Virtualbox VMs/VS2010/C:/Users/George Mauer/.VirtualBox/HardDisks/VS2010.vdi; VD: error VERR_FILE_NOT_FOUND opening image file '/host/Users/George Mauer/Virtualbox VMs/VS2010/C:/Users/George Mauer/.VirtualBox/HardDisks/VS2010.vdi; (VERR_FILE_NOT_FOUND).
You see what's going on? With wubi, the windows drive gets mounted at /host/ but virtualbox is for some reason appending an absolute path! I would very much like to use the same exact VM file since it would retain Snapshots and I would be able to use it in either windows or Ubuntu mode. However, even if I try to simply mount the drives into a new VM I get an error: Failed to open the hard disk /host/Users/George Mauer/.VirtualBox/HardDisks/VS2010.vdi. Cannot register the hard disk '/host/Users/George Mauer/.VirtualBox/HardDisks/VS2010.vdi' {guid...} because a hard disk '/host/Users/George Mauer/VirtualBox VMs/VS2010/C:/Users/George Mauer.VirtualBox/HardDisks/VS2010.vdi with UUID {guid...} already exists.
This is especially odd since this worked fine with my recently created Android VM, though this might have something to do with the fact that VirtualBox recently changed their default VM storage locations. Any idea how to fix this? My Linux-fu is weak but I seem to remember from CS class something about symbolic links that might be relevant here?
On my server I provide OCR file conversion service but the problem is when a user uploads a file and it's being converted then if you open another tab and try to load the site it won't respond until that conversion is completed. In other words until the PHP script finishes execution apace doesn't serve any other request to the same browser.
Here is my apache configuration:
Code:
ServerTokens OS ServerRoot "/etc/httpd" PidFile run/httpd.pid
[code]....
You can check what I mean if you try to upload and convert a file and while the file is converting try opening the site in another tab.
I've had a VPS running Ubuntu 9.10 x64 server, hosting 3 websites of mine for a few months now. This problem has been happening for a while. Every once in a while, probably every 2 or 3 days, I'll wake up in the morning, and apache won't be responding, no web pages will load. /etc/init.d/apache2 status, reports that apache is functioning properly. Every time I simply have to restart the daemon and things run fine for another few days.
I thought maybe it was a memory issue, so I lowered the MaxClients in the prefork module from 50 to 30 a few days ago, but the same thing is still happening. My VPS has 512MB of ram, burstable to 1GB, and according to Virtuozzo, there was only one night of high traffic where I even came close to that soft limit. I've checked my syslog, and there's absolutely nothing in there about apache. I've checked apache's error.log as well, and there's nothing in there that would indicate a problem either.
A Linux (CentOS5.3) server is setup with apache reverse proxy. The reverse proxy server is opened to outside and an internal server is mapped to ProxyPass configuration. SSL certificate is also installed on the Apache reverse proxy server. The problem is, it is extremely slow in serving http requests through reverse proxy. There is no problem with server resources or bandwidth. When the internal server is directly accessed through Internet, there is no delay. The backend server and the reverse proxy server are also on the same switch (same subnet). When I searched the Net, there were recommendations to enable cache in Apache. I did so as follows in httpd.conf.
But still there is no progress. Do I want to enable cache in ssl.conf too? Or is there any other workaround to speed up Apache reverse proxy. Is there a way to check that caching is happening?
Question 1: Hey. I noticed i have quite a few logs that end with .[number] for example "syslog.1" "mail.info.1" etc, why is this and why are they there since almost nothing is logged in them ??
Question 2: on my server im running a script like imagebam and imageshack with hosts images so i have quite a few apache requests to my server. I am wonder why apache takes up so much CPU for some of the requests ? in Htop some requests take up 1.2% CPU while other take up 3-5% etc, so the total load is about 1.50 0.58 0.84 to 2.61 1.08 1.14 with about 128-150 apache requests all the time while sometimes the CPU load can be almost 0 with the same ammount of requests. is this normal? what could cause this in apache ?? the server is just running apache2. MYSQL is running on another server.
I've been scouring the message boards and trying different things for two days and haven't found a solution. I set up Ubuntu 10.04.1 a few days ago using the server iso and selected the LAMP installation option. I will be using it as a sandbox to try out things from a PHP/MySQL book I purchased.
I have installed Gnome since sometimes I just can't get the command prompt to let me do things and Gnome may at least tell me why. I also installed PHPMyadmin.
My problem is that I can work with this thing all I want in any way I want - HTTP, SSH, SFTP, etc. - from within my home network but Apache refuses to respond to an outside request. Since it's only a test box it usually wouldn't be that big of a deal but I am going out of town for a couple of days and would really like to start working with my new book.
I am 99% sure my ISP is not blocking any ports. I have a ComTrend ADSL modem with router. I have given my Apache server a static IP on the NAT (xxx.xxx.x.101) and set up a dyndns address so I don't have to keep remembering the IP address.
Here is what I have tried so far: Setup a virtual server (port forwarding) on the router to to direct traffic on port 80 to internal IP .101The router told me its interface is using 80 so it would move itself to 8080 Made sure to also add port trigger for port 80 Made sure to save/reset the router Used my iPhone to connect via 3G - didn't work using IP or dyndns name Used iPhone to connect to 8080 and router responded
Set port forwarding/triggering for 8080 -> 80. No joy there, either Tried changing listening port to 8000Set port forwarding and triggering to allow port 8000 Changed ports.conf to read NameVirtualHost *:8000 and Listen 8000 Changed first line of /etc/apache2/sites-enabled/000-default to <VirtualHost *:8000> Restarted Apache service
Apache responded to dyndns.com:8000 from home machine No response when trying same on iPhone (sorry, it's the easiest way for me to test from outside my home network)
At one point, I added "ServerName localhost" to the otherwise empty httpd.conf file but that didn't seem to do a darn thing.
Many of the posts I have perused are at least a couple of years old and have included information on taking action with files that are not there or are no longer where they were. I have interpolated where I can but so far nothing has worked.
i use virtual hosts to develop several web applications. These are located in my home folder under /home/user/projects/project After a fresh installation, i always get a 403 forbidden error. After googling and reading on this forum, several solutions are mentioned for this problem. But i can hardly believe putting using a chmod 755 on my home folder is a correct solution. What is the correct way of doing things in this situation?
my server into the public internet and must have my webserver [2.2.9 (Debian)] act secure. But this does not look very easy [I am searching, reading and working on it already the whole day ]. I read the apache docs, but there is a lot of stuff, which is different in Debian [lenny, 5.0.6]. Apache ignores the host-header given by the browser: [URL] are all served, but should be blocked. I new new to apachy, but on my IIS this works as expected. All browsers act equal [so no browser header problem].I configured two VirtualHosts, an excerpt:
The I go to my hosts file on any box, and add hugo's ip-address under the new name x. Then, x is served, although the host-header in apache Every user coming from the internet could make the same!
I noticed i have quite a few logs that end with .[number] for example "syslog.1" "mail.info.1" etc, why is this and why are they there since almost nothing is logged in them ??
Question 2: on my server im running a script like imagebam and imageshack with hosts images so i have quite a few apache requests to my server. I am wonder why apache takes up so much CPU for some of the requests? in Htop some requests take up 1.2% CPU while other take up 3-5% etc, so the total load is about 1.50 0.58 0.84 to 2.61 1.08 1.14 with about 128-150 apache requests all the time while sometimes the CPU load can be almost 0 with the same ammount of requests. is this normal? what could cause this in apache ?? the server is just running apache2. MYSQL is running on another server.
I'm trying to figure out how I can get a request count per CIDR/24 from an apache log in combined format - e.g.: Code: 24.147.44.122 - frank [10/Oct/2011:13:55:36 -0700] "GET /apache_pb.gif HTTP/1.0" 200 2326 "[URL] "Mozilla/4.08 [en] (Win98; I ;Nav)"
I'm stuck using BASH for this and I generally write everything in Python, or even Perl if necessary so BASH isn't the most comfortable for me. I've got enough to extract the IPs and get a count I just need a slick way to come up with a sum on a per CIDR basis.
I'm running wordpress and have gone to the painstaking effort of setting up the ftp server on my 5.3 machine so it can do its own updates and download plugins. However, I've found that if I try to download anything, I get a "unable to resolve host..." error from the script. If I watch Iptraf while making the request, I see all the port 53 requests going to 127.0.0.1. Pings are too.
However, if I ssh in, I can ping and wget and whatever I want all day long and all the DNS requests to go the router (Clarkconnect 3.2 gateway machine). The web browser in the terminal works fine too. What is it about the php scripts that is causing them to misdirect pings and dns requests? I've never seen anything like it and can't find anything on the web about it either.
My test messages from one mail server to another are getting sent to my spambucket and the only error in the logs is: warning: numeric domain name in resource data of MX record for [URL]: 74.63.64.42
after installing Debian 6 on a server. When I try to install a software called ActiveCampaign, I get the following error ... "Your server does not appear to be handling sessions properly."I have install Apache, PHP, MySql and Perl already. Also, after the server restarts, I Webmin will not automatically start, even thou it is setup in the Webmin configuration to start with the server. I have to use /etc/init.d/webmin start from a command line after I su.
My last question is about ftp permissions. I have install proftpd and it seems to be working fine, but when I try to edit any file or upload, I can not. In order to upload and manipulate files, I am using WinSCP under root, wich is a big NO NO.Sorry for the three questions, but I figured I would ask all in one post, instead of creating multiple, since I already have your attention.
This has been driving me absolutely crazy for a few days now. I have an e-mail. If I send it from a Windows Server 2003 server via MailEnable, a certain person receives it using Outlook 2003 just fine. If, however, I send the exact same message from an Ubuntu 10 server using SSMTP to relay the e-mail to the exact same MailEnable queue, the recipient receives a messed-up-looking e-mail. There are all kinds of "= " all over the place. The exact same e-mail looks fine on my Ubuntu computer, as well as my Mac, and all of the Windows/Mac computers here at work. Here is an excerpt from the copy they receive:
COMPANY has been communicating with COMPANY as they Agency worked to develop the= COMPANY permit, knowing that this would impact aquatic pesticide applicat= ion. But in December, COMPANY advised us that the COMPANY permit would = also cover any pesticide applications made "on or near any water or water c= onveyance." They described a scenario in which a pesticide applicatio= n made in a field or pasture that had a waterway or ditch in it (even if th= e waterway or ditch was dry) would need a COMPANY permit. This meant th= at the majority of pesticide applications made to crops would need a COMPANY = permit, and would have to comply with all the burdensome paperwork requirem= ents plus all of the pesticide application records would be available to th= e public for review.
What on Earth is happening? I'm using PHP's Mail_Mime to send the mail, if that matters.
I have tried testing this on an in-house netbook, running XP and Outlook 2003, same as them, and it looks fine. Unfortunately though, it's happening for more than one person, with different companies, in different states. Is it maybe an Outlook thing? Or a firewall?
Here are the headers they sent me, when I requested them earlier:
Squid document says that Squid accepts only HTTP requests but speaks FTP on the server side when FTP object are requested.
We call Squid HTTP and FTP caching proxy server. Does it also caches FTP contents? Is it possible to configure FTP clients to use Squid cache? When we make an FTP request to an FTP site via Squid will it be bypassed?
I have a home DNS server that has been working for some time today. Today I restarted to restarted it to clear the cache on it and now it refuses to answer and requests. Named starts fine with no errors. Here is named config file that worked for about 2 weeks fine and now doesn't want to work.
I am trying to solve a problem where Apache stats aren't displaying correctly in Munin. I've ran through quite a bit of checks and tests regarding Munin setup, but I think my issue is related to Apache, but my skill set there is lacking.
first, system info: monitored server: CentOS 5.3 2.6.18-128.1.1.el5
i'd like to have logrotate compress the logs that are older than 3 days. Is this possible with logrotate, or do i just schedule a cron job to bzip everything under the folder older than 3 days?
I'm trying to forward VNC requests from server a to server b, acctually I need the server a to be just VNC proxy and other servers behind server a can be responsible for VNC requests.I did it with this iptables rule but it didn't workAnd for notification all VNC sessions of mine are in range of 59100 to 59199
I'm just asking about a script (ex, bash script) that will let me know how many requests to each website on the server? So is there a way to get know from shell how many requests or connections to each web site on the server, in order to determine which website is under flood or DoS/DDoS attack.
I have an nfs server that is getting tons of mount requests, and i suspect it's becoming a performance issue. Even when noone is in the office I see mount/unmount requests in my logs. Is this normal?
/var/log/messages: Mar 30 10:32:42 morgan mountd[29643]: authenticated unmount request from uranus:757 for /morgan/users (/morgan/users)
Background:I have a small PC104 running opensuse 11.1. I'm writing a small client/server application for debugging purposes using mono and WCF. All the client does is make a request for information every 100ms.Problem:After about 20 requests the server quits responding to the client. Even if the client is running on the same machine. I've run the exact code on another laptop running opensuse as well as a laptop running windows and everything works great. Hopefully that closes the option that it is a code, mono, or opensuse flaw.Is there a kernel option or a network option that anyone knows about that might cause this sort of behavior?
We have a cisco ASA firewall at work,which redirects all http traffic to our webserver. We have to install a new website ,but it can't be installed to the same server. Setting up a squid reverse proxy can redirect the incoming http requests to the appropriate webserver? If yes, could I get some directions on howto?
I run a small home server (Debian 4), which acts as my gateway to the internet (ie, firewall) and runs a web server, dhcp, dns, and acts as a file server to the rest of the machines on my home network. Now I know it's never a smart idea to have all those services running on the same machine that is acting as a firewall, but I don't fancy running multiple servers just for home use, as it's mainly allowing me to learn system administration.
I noticed a few days ago that my internet had become unbearably slow, to the point where I could sometimes not load web pages. I spent a while searching through log files on my gateway, to try and find out what was eating up all of my bandwidth. When I came to apache's access.log file, I was confronted with this:
Multiple requests to my server, for totally random websites. I didn't even know it was possible to make those types of queries to a webserver. The only thing that is on the web server is a browser based torrent client. I have only shown a small snippet of the log file, but there are around 90k lines to different web addresses, from many different IPs. What I want to know, is what is happening? :S Why is someone querying MY web server, for web sites totally unrelated to it? And most of all, how can I stop it. My initial was to try and use iptables to block multiple requests from the same ip within a certain time frame, which I think would work as the server shouldn't really get many queries from external networks.
