steviewdr

http://goosh.org/ Pretty neat. Basically its a cli gui to google. -steve

http://www.debian.org/doc/manuals/securing-debian-howto/ The concepts there will apply to most linux systems. -steve

grep "keyword" access.log > fileout.txt -steve

I'm very sorry. I missed one key line in my above config. Edit your php.ini and add the following: auto_prepend_file ="/etc/set_php_headers.php" Then /etc/set_php_headers.php has: <?php putenv("HTTP_HOST=". $_SERVER["HTTP_HOST"]); putenv("SCRIPT_NAME=". $_SERVER["SCRIPT_NAME"]); putenv("REMOTE_ADDR=". $_SERVER["REMOTE_ADDR"]); ?> Job done. Jab me here again if it still doesnt work. -steve

Yes, I got my wires crossed badly there. chmod 2775 /var/www or chmod g+s /var/www I'm not so sure about the umask of 0002 been default. I think it could be 0022 for debian and ubuntu. -steve

Try using the chmod +t option so that any files created by anyone in /var/www are automatically created 775. Note: not recommended. Use suPHP or suExec to have apache change owner etc. See: http://wiki.kartbuilding.net/index.php/Chmod#Make_a_Sticky_Directory_or_File -steve

Nah. All webapps need to interface with an imap server in order to read and write emails to mbox. Installing imap and/or imaps is easy nuf with courier. http://wiki.kartbuilding.net/index.php/Courier_IMAPS_Server_-_Maildir Courier does use ~/Maildir instead of ~/mbox -steve

This works fine in a production environment: Edit your php.ini and change to the following: sendmail_path ="/usr/sbin/sendmail_wrapper" Then create the following file: /usr/sbin/sendmail_wrapper with this in it: #!/bin/sh logger -p mail.info vhostmail: site=${HTTP_HOST}, client=${REMOTE_ADDR}, script=${SCRIPT_NAME} /usr/sbin/sendmail -t -i $* Chmod 755 /usr/sbin/sendmail_wrapper Your off. Tail /var/log/mail.info for a log of email sent via php's mail(). -steve

Do a netstat -tap. It should have: tcp 0 0 *:smtp *:* LISTEN - (if there is a localhost/127.0.0.1:smtp, its bound to localhost only) A lot of ISPs block port 25 to prevent spam from been sent via hijacked pcs. Its strange to see incoming traffic on port 25 to be blocked however. -steve

Ive used apache + mod_proxy to access various tomcat applications, and never had to wait 3 minutes for an initial connection. One thing it could be - your apache box is trying to do a Reverse DNS lookup on the ip accessing it. Generally rdns lookups can take up to 3 minutes before they timeout. An easy way to test whether its an internet /network traffic issue would be to install httping and have it run ever 30 minutes overnight etc. when you know that network speed is not an issue. Lastly, the checkpoint firewall, you might have to start logging on it to see what packets are going/coming during that 3 minutes. -steve

There are so many variables, you will have to debug it yourself. Start with the tomcat 6 box. Put elinks or firefox on that machine and go directly to tomcat and do your tests. Lather, Rinse, Repeat for all levels. As for having 2 different apache servers doing nothing but mod_proxy, this sounds very excessive. -steve

http://knopper.net/knoppix-mirrors/index-en.html http://www.ubuntu.com/getubuntu/download http://fedoraproject.org/get-fedora.html (Live CD option ^ ) As for best linux distro for programmer, I would choose ubuntu. -steve

Format and reload the OS. Keep all security updates for the OS applied. apt-get update apt-get upgrade -steve

apt-get install vncserver You only have to fire up X and VNC when you want to use the GUI on the remote server. It works quite well over broadband. -steve

Well if you trust the few users on that system your fine. What they could potentially do is: Create a php/cgi script which will copy all /var/www/htdocs and allow them to download the source code and see your mysql passwords. As php scripts run by the webuser, i.e. apache, then can access your files. Ideally if you trust/work with the other people, you might not mind too much. You could also disable php and cgi (and other scripts) by default for all users. So they can only run static html files etc. You could look into suPHP which is on apt also. It allows the webserver to "su" to the owner of the folder with the php file been accessed is in. -steve

wow. apt-get autoremove seems to be in ubuntu, and not debian (etch). If you want to reinstall all them packages, you need to go: apt-get install ........<insert list from first post here>.......... -steve

Try %1 instead of %2. -steve

Is it a dedicated server? Are there any other users of the same box/server? If not, then it will be fine to chown -r apache2:apache2 /var/www/htdocs/live (and chmod 600) after that. TBH, chmod 777 and the like only comes into effect if there are multiple users on a server, and or if there are some scripts which are poorly written. </two-cents> -steve

Although using "aptitude" is the normal/typical way to install new debian/ubuntu packages, I only use apt-get with a few dpkg commands aswell. -steve

While openssl and sshd are fixed in recent debian and ubuntu updates (apt-get upgrade), it is *not* this simple! All ssh and ssl keys which were created on debian or ubuntu servers between September 2006 and last week, need to be recreated. If you use ssh keys on debian/ubuntu, these need to be deleted, and new ones created. After a couple of apt-get upgrades in debian etch, a new tool called "ssh-vulnkey" is installed. Simple run "ssh-vulnkey -a" as root to detect if ssh keys are at risk. Its quite a tale and there has been a lot of noise about it. See: http://wiki.debian.org/SSLkeys#head-5450db0076b3d85650f72117a9884f89d2349032 http://mail.linux.ie/pipermail/ilug/2008-May/097975.html -steve

I think its a ssl cert issue. Its as if that active x requires a registered ssl cert. Not sure. You'll have to google for: http://www.google.ie/search?hl=en&q=https+error+reason%28134%29&meta= -steve

Look closer at your error logs. Get webalizer or another program to give you statistical output. Uncomment server-info from your httpd/apache.conf and enable the module server-info. How much memory is in your server? Whats your worker/process thread settings in your httpd.conf/apache.conf. -steve

I dont think you will be doing this quite easily. Generally one smtp server does all the delivering of emails. Your best bet would be to set the reverse dns of your main ip to be something like: mail.hostingco.com that way when you look at the email headers it will look official. It might be possiblel to do what you are saying, but I wouldnt think so. You could use vserver or a virtual machine to handle the final delivery of mail, but that would involve a lot of setup and management. -steve

If you go to "Add remove windows components" in the add remove programs in the Control panel, you will be able to enable a mailserver and iss if you want. Apart from that you'll have to download some sendmail or smpt program for windows. -steve

You might be aswell to do a: yum remove dhcp (do a "yum list installed | grep dhcp" first to see the exact name of the dhcp package installed). Do you have just the one network card? (try doing: "lspci | grep -i ethernet" to find out. lspci might not work by default on fedora). It could well be that your gui sits ontop of the sysconfig and changes it after the system boots. I would try removing the dhcp package and then seeing. -steve