Comments on The Perfect Setup - CentOS 4.4 (32-bit)

The Perfect Setup - CentOS 4.4 (32-bit) This is a detailed description about how to set up a CentOS 4.4 based server that offers all services needed by ISPs and hosters (web server (SSL-capable), mail server (with SMTP-AUTH and TLS!), DNS server, FTP server, MySQL server, POP3/IMAP, Quota, Firewall, etc.). This tutorial is written for the 32-bit version of CentOS 4.4, but should apply to the 64-bit version with very little modifications as well.

9 Comment(s)

Add comment

Please register in our forum first to comment.

Comments

By: hoihtah

Thank you guys for putting up this well written guide.

 Just one question,  how do I do this setup with mysql version 5 instead of 4?

By: orentocy

Enable CentOS plus yum repository in your /etc/yum.repos.d/CentOS-Base.repo, then you will be able to upgrade both your mysql and php to version 5.

By:

Hi, I enabled cetosplus section using enabled=1,  No my system is updated with php 5 and mysql 5 with the command

yum update -y 

Enabling the centosplus section: 

vi /etc/yum.repos.d/CentOS-Base.repo
[centosplus]
gpgcheck=1
enabled=1

By:

there is one more thing you need to do.  update php.conf file

 cp /etc/httpd/conf.d/php.conf.rpmnew /etc/httpd/conf.d/php.conf

Otherwise, httpd will error out when trying to start.  Or at least it does on mine.  :) 

By:

Hi all,

I found some bugs if the yum CentOS Plus is enabled before starting the ISPconfig  OS  preparation.

If it happens to you, go back to mysql4 and php4, make your ISPconfig prep THEN enable the CentOS plus repo to install Mysql5 and php5.

Now you are ready for ISPconfig install.

Thanks for this perfect howto. Saves a lot of hours. 

By: jperrin

Very good tutorial, and very detailed, however one part concerns me. Your rebuild of zlib at the end does not address removing the currently installed zlib, or address the problem of future rpms which may rely on zlib failing because of the one built from source (rpms are rather ignorant about source built software). I would also posit that you cannot rely on the version of zlib to identify that it's vulnerable. Security fixes are backported in centos (and it's parent distro, RHEL), so version numbers may be inaccurate. The changelog for the zlib rpm lists several CAN- advisory fixes, so I wonder if the bug you claim is one of these. If it is not, has this been reported to the centos folks, or to the upstream RedHat bugzilla?

 If this bug is not fixed in the RPM as one of the listed CAN changes in the changelog and the rpm does indeed contain vulnerable code, I'd like to see it fixed in the distro, rather than being bolted onto a(n excellent) tutorial.

By: till

I dont think that theare is really a bug in the zlib that ships with CentOS, the problem is that the version number dont get updated when the fixes where applied.

For example if you want o compile ClamAV which is nescessary for ISPConfig, Clamav complains about a bug in zlib and stops compiling. So either the ClamAV team has to add a better zlib detection routine or the CentOS team has to set a higher version number in the zlib library when they apply fixes.

By: jperrin

This is addressed a bit more thoroughly in the post by Johnny Hughes, who is one of the CentOS Project leads, https://www.howtoforge.com/perfect_setup_centos_4.4_p6#comment-3055 What it comes down to is an upstream versioning decision by redhat, which centos inherits as a clone/rebuild product. I would consider this to be a flaw in ClamAV/ISPConfig packaging, and that it should not be advertised as a CentOS vulnerability unless such a problem actually exists.

By:

First, thanks for an excellent tutorial!

I had serious problems with ntp running the Perfect setup on a Windows host using VMWare GSX server. My clock was constantly running behind and I would use rdate to set the clock but very soon the clock was running behind again.

Googling I found a workaround that worked out well (if running SMP on single core processor):

1. Edit /etc/grub.conf
Add 'noapic nosmp nolapic clock=pit acpi=no' so your grub.conf looks like this:

title CentOS (2.6.9-42.0.10.ELsmp)
        root (hd0,0)
        kernel /vmlinuz-2.6.9-42.0.10.ELsmp ro root=/dev/VolGroup00/LogVol00 noapic nosmp nolapic clock=pit acpi=no
        initrd /initrd-2.6.9-42.0.10.ELsmp.img


2. Edit /etc/ntp.conf
Add 'burst iburst' after your server:

# --- OUR TIMESERVERS -----
server 0.pool.ntp.org burst iburst
server 1.pool.ntp.org burst iburst
server 2.pool.ntp.org burst iburst

This solved all my problems with a slow clock and my time is now on the spot 24/7.

My Windows system:
P4 3 GHz
3.5 GB RAM
VMWare GSX server