sameold's questions

I have a number of domains that I'm hosting on my web server, so I've edited their A record to point to my server's IP address. Occasionally I have to change the server's IP address, which means I have to edit the A record for each domain hosted on that server. There's also risk that a website would go down because the change was not made for its A record.

I'd like to simplify this if possible. Is there a way I could set the A record to a domain or subdomain, then set that domain/subdomain to the desired IP, so that when the IP changes, I would need to change only 1 location instead of multiple domains? I'm open to ideas. It does not have to be exactly this way through the A record. Anything that gets me the desired effect is fine.

When you register a domain, typically the registrar will set it up on their default DNS servers and point it to their default web server to serve their default parked page.

Obviously the web server (or hosting company) serving out the pages will always know traffic, so in the Registrar-DNS-Host chain, I'm really asking about the first 2: registrar and DNS provider.

If you edit the zone file with the registrar to point to your own web server, the website is no longer hosted with the registrar, but the registrar still hosts the DNS. Would hosting DNS still enable them to count traffic?

If you change nameservers and point it to a specialist provider like DNS Made Easy, then the registrar has nothing left (except basic registrar functionality, like WHOIS, etc.) So I'm guessing at this point, they have no way of estimating traffic at all. Is this correct?

I'm basically trying to verify my guess that only the Host or DNS provider can count traffic, but the registrar itself has no way of doing that.

I'm using mod_write to rewrite this

www.variabledomain.variableext

to

http://my.com/variabledomain.variableext 

Note that variabledomain and variableext are really variable, so I can't hardcode them.

I'm not an expert at mod_rewrite, but I thought something like would work, but it isn't. Any ideas what I should be doing instead.

RewriteRule ^(.*)\.(.*)\.(.*)$ http://my.com/$2\.$3 [R=301,L]

I'm testing a postfix installation. I noticed that if I try to send mail from a fictitious account on another domain that isn't mine ([email protected]), to one of my local users ([email protected]), the email still gets delivered to my user.

MAIL FROM:<[email protected]>

RCPT TO:<[email protected]>

I'm not sure if this is normal behaviour or if it can cause problems. For example, I'm worried that users could use the server to send spoofed emails that aren't really originating from the email accounts they claim to be originating from.

P.S. I'm still in the early stages of learning postfix, so not sure if my question makes sense.

As per @DerfK's comment, I tested for an open relay (where both MAIL FROM and RCPT TO aren't hosted on this server), and this is what I get. For some reason it accepts the RCPT TO, then says goodbye AFTER I type the subject. So is this good or bad?

HELO test.mydomain.com
250 mail
MAIL FROM:<[email protected]>
250 2.1.0 Ok
RCPT TO:<[email protected]>
250 2.1.5 Ok
Subject: testing for open relay
221 2.7.0 Error: I can break rules, too. Goodbye.
Connection closed by foreign host.

I'm trying to understand how all the components of a mail server fit together.

• I understand that the SMTP server is necessary to send emails.
• In terms of accessing email, I'll setup a webmail client but do not need access through a desktop client.

So my question is: do I need an IMAP server in this case? or can the webmail client somehow link directly to the SMTP to send and read emails?

I'm told that it's best practice to have a secondary MX record pointing to a secondary mail server in case the main one fails.

If the main mail server fails, how will the users on the main server get their email? Do I need to mirror all user accounts on the secondary server?

I'm new to mail servers, so I'm not sure if my question makes sense.

I'm using Amazon Linux AMI, which seems to be some RHEL or CentOS offshoot. I've read that it should be binary compatible with RHEL. My question is: is there any advantage to compile a package from source as opposed to using the RHEL or CentOS repos (as they should work too)?

I have a .rpm file on my windows machine. Is there a way to inspect this file (either on windows or centos) without installing it, to find out where it was compiled from? something like metadata?

I'm new to mail servers and still trying to understand all the components.

What I read is that there are many MTAs, but the 4 common ones are qmail, Postfix, Sendmail, and Exim.

I also found this list of mail servers. http://en.wikipedia.org/wiki/List_of_mail_servers

What's confusing me is that mail servers like Zimbra and Atmail are listed at the same level as qmail/Postfix, etc. Are mail servers and MTAs the same thing? I thought (correct me if wrong) that a mail server includes an MTA as one of its components.

In several CentOS tutorials, they advise enabling the RPMRepository. I have some questions about this repository.

1. It seems like it's a RedHat repository that CentOS uses as well, since it's based on RedHat. Is this correct.

2. How do I read the content of this rpm i.e. what packages are inside this repository. I've downloaded one of the rpm files at http://pkgs.repoforge.org/rpmforge-release/ and tried opening it in a text editor on windows, but it's not a text file. How do I know what list of packages I can get from this repository?

3. Is this an essential repository? What I mean is: are the packages available from this repository unavailable in any other repository? They're only available from this repo?

4. Are there other repositories that are of use for a system running CentOS?

When installing CentOS (6.2), it installs a whole bunch of packages, but the installation is often very fast, it's hard to note the names of the packages. I have a couple of questions:

• 1) I'm guessing I could wait for the entire installation to complete and then query for the list of installed packages on the system. How do I do that?
• 2) Is there a printed list somewhere on the web, so I don't have to install the OS first just to get the list?

I often hear people making statements such as "our MySQL server machine failed", which gives me the impression that they dedicate a single machine as their MySQL server (I guess they just install the OS and only MySQL on it). As a developer not a sysadmin, I'm used to MySQL being installed as part of a LAMP stack together with the web server and the PHP.

Can someone explain to me:

• what's the point of installing MySQL on a separate server? sounds like a waste of resources when I can add the entire lamp stack there and additional servers as well.
• if the database is on a separate machine, how do the apps that need to use connect to it?

I'm looking into virtualization where I'll set up a few virtual machines on a single physical hardware (similar to what vps providers do). Each of these VPS will mainly be used as a webserver.

I'd like to make a PHP framework and certain PHP libraries available to each of these vps, but would prefer not to have to store the PHP libraries on each vps. I'm concerned about initial setup of each framework/library (copying the same files to multiple vps), and updating the libs/frameworks (performing the same update for each vps).

So my question is: given the nature of the files (they're PHP frameworks on libraries), what are some of the ways I could share these files between VPS? and is there a way that wouldn't reduce efficiency? in other words, the files would be accessible just to the vps as if they're residing on it.

I have a question about type 1 and type 2 hypervisors. Type 1 runs on the bare metal, and then the end operating systems are installed on top. This seems to me like the most logical way to build a virtualized environment.

But there's type 2 where a base OS has to be installed first, then the hypervisor, then the end operating systems needed. I don't understand why anyone would go through the trouble of using a type 2. Are there any use cases that justify needing a base OS as it's done in type 2?

I've installed ubuntu server on a VirtualBox virtual machine and now I'm trying to follow a tutorial that says I need to install a package called csh. When I type this line

sudo apt-get install csh

I get "unable to locate package csh". This happens with all 4 other packages I tried to install. I'm not sure what I'm doing wrong. Perhaps the VM can't reach the Internet. How do I check, or troubleshoot this?

I have a host machine that has virtualBox installed, and a guest OS virtual machine is created on it.

I want to test installing virtualBox on that guest OS for training purposes, so this means that I'll be installing VirtualBox on a machine that's actually a guest image running using VirtualBox itself. Do you see any problems with this?

I have a question for SysAdmins who know many flavors of Linux. I need to be familiar with SUSE Enterprise for a job, but I currently have no access to the enterprise edition. I also would prefer to learn Ubuntu. So how similar is ubuntu to SUSE Enterprise in terms of the commands, etc.

I'm reading the requirements for a virtualized environment from here which says:

IP Address

- One IP address on a subnet for the host.
- One IP address on a subnet for each VM Guest.

I want to know if a normal home network (router connected to the Internet) would meet this requirement?

I used virtual box before to test OS installation on my own personal computer. I'm not sure if it's suitable for actual virtualization on live servers. I mostly hear about Xen for that. Is virtualbox at the same level (production environment virtualization)? and what other alternatives are there, other than Xen?

I have a dedicated server. I read here that VPS is better because it can be booted and restarted quickly. I'm not sure if this applies to my case or not, but my question is:

Is it better to have 1 VPS that's setup to take all full resources of the machine than to install the OS directly on the full resources of the machine without virtualization?