Tag Info

Hot answers tagged

32

Install dpkg-dev: sudo apt-get install dpkg-dev Add repository: sudo add-apt-repository ppa:nginx/stable Edit /etc/apt/sources.list.d/nginx-stable-lucid.list, add dpkg-src: deb http://ppa.launchpad.net/nginx/stable/ubuntu lucid main deb-src http://ppa.launchpad.net/nginx/stable/ubuntu lucid main note: (the previous step may have already been ...


11

Usually it is something like that. ./configure This (when executed in the directory your source code is), creates a Makefile with the rules that will be used to build the program. This is where you can change the behavior of the compilation (that is, compile the program with options that are not default). make This builds the program with the rules ...


10

Well, the closest thing you can do (that I'm aware of) is to query the OPTFLAGS variable of the exim package: [root@fedora11 ~]# rpm -q --queryformat="%{NAME}: %{OPTFLAGS}\n" exim exim: -O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic [root@fedora11 ~]# You'll get better answer, ...


10

Welcome to Server Fault. In general this type of question is not on-topic here. It is expected that most sysadmins will have already faced this issue before they became a sysadmin through tinkering with their own workstations. However, sysadmins coming from a Windows background may not have faced this, so here are some things to consider: Why not use a ...


8

I can't disagree with Danila -- if you have a support contract, use it. In fact, you'll probably want to use it to upgrade from Solaris 11 to Solaris 11.1. That said, your problem is that the local repository you've attached is newer than your installed system, and doesn't have the matched set of packages. You want to get the Solaris 11 repository ISO ...


7

These should be installable via yum and the base RHEL repositories. Have you tried querying if they are already installed but not in your path? Also, have you successfully installed anything via yum? To check what package you need to install: [user@server]# yum whatprovides autoconf automake Loaded plugins: rhnplugin, security autoconf-2.59-12.noarch : A ...


7

Generally I keep up-to-date using the vendor latest packaged version. In the case of RH (and Centos) although their major release number might not change on a package, they usually backport key and critical patches from the later releases. To make things easier for me, in the event that I do really want to go to a much later release in a package is to ...


6

In the latest version of PHP's configure.in at the time of this answer, --with-layout is defined as: PHP_ARG_WITH(layout, layout of installed files, [ --with-layout=TYPE Set how installed files will be laid out. Type can be either PHP or GNU [PHP]], PHP, no) case $PHP_LAYOUT in GNU) oldstyleextdir=no ;; *) ...


5

You have ten Lenny servers all running the x86 arch. It's completely possible to build a single package that installs to any computer running a specific arch. It's good practice to build Debian packages with pbuilder, to avoid any "works on the dev box" problems. It's relatively simple to rebuild a package. apt-get has a source command that will pull down ...


5

Since you are doing this in a script, you could alternatively check the exit code of the commands as you run with the $? variable. You could do something like: ./configure if [ $? -ne 0 ] then echo Configure failed exit 1 fi make if [ $? -ne 0 ] then echo make failed exit 1 fi make install if [ $? -ne 0 ] then echo make install ...


5

I'm not going to blow up my workstation for days on end by testing this, but given a directory of SRPMS I don't see why this wouldn't work: time (for RPM in $(ls *.srpm); do rpmbuild --rebuild $RPM; done)


5

If you look inside the default Ubuntu /etc/nginx/nginx.conf you ought to find the following line. http { ... include /etc/nginx/sites-enabled/*; } That is what you need to get the sites-enabled magic happen.


4

I found that I was able to get this working by changing the rfc2440.c source itself. Since my malloc.h file was located in /usr/include/malloc/malloc.h, I changed the line from #include < malloc.h> to #include < malloc/malloc.h> and compiled. There were no compiler errors, and it installed fine with make install afterward (technically, sudo make ...


4

At command line: wget [link to php goes here] tar -xvf [file you downloaded goes here] cd [name of file without the extension\folder that was just extracted to] ./configure --enable-zip make make install In convenient batch form that you should be able to copy and paste: yum install -y make wget gcc zip-devel bzip2-devel wget ...


4

I tend to prefer to standardize my system builds using primarily scripting and source. Others often like to use configuration management tools and creating their own packages from source for their distributions native package manager. If you standardize your build properly, these methods largely provide the same results and they are simply different ways ...


4

Since squeeze in Debian and since oneric in Ubuntu, there are some (like headers-more-module I needed) in apt-get install nginx-extras


4

Remove the commented lines from your ./configure list. They are causing --enable-mime-magic to be processed as a separate command.


4

I use the packages version, ESPECIALLY for internet-facing applications. If you use a major distro like Centos, Red Hat, Ubuntu, or Debian, you will get all the latest security fixes, even if it is not the latest major version number. Security and stability are much more important than having the absolute latest features, especially with something like ...


4

Create a build environment and and set up a long chain of rpmbuild --rebuild packagename.srpm commands... But the better* approach is to simply go to CentOS.org. Download the ISO images for the current release from a local mirror server. *Assuming you're not modifying any packages.


4

Since you asked for best practices: The best one is to don't install software not in the repositories. If that's not possible, the best practice is to build packages out of it and use the systems package manager to manage them. You will need to come up with a way to check for important updates, but how you do that is up to you. But if they are there, you ...


4

Sorry for my english first. my os: CentOS 6.3 mini without desktop x86_64, i haven't yum install libevent i met same problem many times in different environment Here is my installation process of libevent: tar zxvf libevent-2.0.20-stable.tar.gz cd libevent-2.0.20-stable ./configure --prefix=/usr/local/libevent make make install before my install ...


4

nginx -s reload is not sufficient to upgrade to a new binary. Read this entry on the wiki to see the series of signals that need to be sent to upgrade to a new binary. Alternatively, since you're already installing from source, there's a make upgrade target you can run after make install that will send the signals for you.


4

root@ip-10-112-39-42:~# apt-cache search geoip .... libgeoip-dev - Development files for the GeoIP library .... root@ip-10-112-39-42:~# apt-get install libgeoip-dev


4

If this is production SPARC T4 box with Solaris 11 on it. If i was in your situation, I would not mess with it and get a hold of Oracle support. Especially in the case of the doubt while installing one of the crucial systems aspects as system headers for "gcc" while being in the state of "I'm not at all familiar with how solaris package management works." ...


4

The file include/generated/autoconf.h is generated in the make prepare step. If you are trying to build a kernel module, you will also need the make scripts step: gunzip < /proc/config.gz > .config make oldconfig make prepare make scripts Usually the kernel is accompied with a headers package, have you tried installing that first? For CentOS, try ...


3

What I get the source for an application the first thing I do is to read the pretty standard README/INSTALL file. They usually tell me exactly what I need to do. They tell me what dependencies I will to install before I compile to get all the functionality I want. They tell me what I need to do to secure the install. less INSTALL less README Take 3-5 ...


3

If you're stuck with compiling yourself, gcc accepts the -m32 option which should instruct the compiler to generate 32bit objects/executables. (Although, if you can go the route gekkz suggests, do that).


3

The best option (IMHO) is to set all the options at the very start. You can do this by running make config-recursive.


3

configure deals with differences between the libs and tools installed on the target system - see this page for an overview - and generates a Makefile (usually a Makefile in each sub dir too). A Makefile is a script usually defining several types of operation - you select the operation by specifying it as an argument to 'make'. 'make clean' (by convention) ...


3

I'm going to go against the flow on this one. For something that faces the Internet I prefer to use the latest stable version, so I build from source. I don't know about Red Hat but CentOS packages tend to be a fair bit behind the latest versions, which means you may well be installing something that already has known security issues. I just get nervous ...



Only top voted, non community-wiki answers of a minimum length are eligible