John R.'s questions

Using a log monitoring tool I had noticed an IP had attempted 100-200 logins as root though failed (PAM's logs caught that), those IPs differ slightly but originate to one single domain which is a dedicated server host somewhere which I do not care if I block access, no loss.

(EDIT: Or maybe an IP range if DNS resolution is not done during checking.. I do not know how to do that either.)

I added the IPs to my iptables, how would I block a whole host (foobar.com) from accessing any of the services on my VPS including SSH? I am using Ubuntu 10.04LTS Server.

The proxy works fine, my distro is Ubuntu 10.04LTS (Server) and rpaf is loaded within apache, with the following settings httpd.conf:

<IfModule mod_rpaf.c>
  RPAFenable On
  RPAFsethostname On
  RPAFproxy_ips 127.0.0.1
</IfModule>

I can be sure it is loaded, for example:

#a2enmod rpaf
Module rpaf already enabled

I had force-reloaded apache, and reloaded nginx but all my connections are still from 127.0.0.1:

127.0.0.1 - - [13/Dec/2010:23:56:27 -0800] "GET /index.php HTTP/1.0" 200 1267 "-" "Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.0.19) Gecko/2010031422 Firefox/3.0.19"

Why is it still localhost from external? I'll give any info if needed.

I set up munin to run as root for testing, my relevant config is as so:

munin.conf:

[localhost]
    address 127.0.0.1
    use_node_name yes

I tried fqnd.mysite.com (but it couldn't connect), localhost.localdomain too which runs but I'll explain in a sec

munin-node.conf:

allow ^127\.0\.0\.1$

# Which address to bind to;
host 127.0.0.1
host_name localhost
port 4949

Now it all works fine if I type munin-run cpu and it lists myspecs, but when I telnet to localhost:4949 a session will be like this:

# munin node at localhost
list
cpu df df_inode entropy exim_mailstats forks fw_conntrack fw_forwarded_local fw_packets http_loadtime if_err_eth0 if_eth0 interrupts iostat iostat_ios irqstats load memory munin_stats nfs4_client nfsd nfsd4 ntp_kernel_err ntp_kernel_pll_freq ntp_kernel_pll_off ntp_offset open_files open_inodes postfix_mailqueue postfix_mailvolume proc_pri processes swap threads uptime users vmstat
fetch df
.
fetch cpu
.

As you see plugins are listed! But for some reason it replies with a '.', I am running this under root so I don't think it is a permission problem. Any clues will get me off my computer at 4:32AM!

I finally got postfix up and running for the first time, and want to make sure all things are correct before I 'test' it out, so I can assume unworldly powers stopped it from working and not me.

Assume my FQDN is me.mysite.com, should my MX record point there or to mysite.com?

I am on Ubuntu 10.04 and had postfix installed, I can send mail with sendmail but when I send mail to my server through hotmail or alike it just never appears.

I installed mailutils and ran mail while in root (which is who is configured to receive everything) and it said no mail. There was also no mail in /var/mail/root

What steps would I need to do for postfix to start accepting e-mail externally, any specific guides (I prefer not to do virtual users if that is too complex..)

I get this error when checking my DNS with an external tool:

Spam recognition software and RFC821 4.3 (also RFC2821 4.3.1) state that the hostname given in the SMTP greeting MUST have an A record pointing back to the same server

It essentially states this path:

mail.xxxx.com. -> mail.xxxx.com -> (IP of main server) -> xxxx.com

Now my postfix greeting appears as so:

$myhostname ESMTP $mail_name

Is the problem because $myhostname is mail.xxxx.com and that is re-pointing to mail.xxx.com and THEN to my main server? I think I should change it to $mydomain

I would like to get some insight before I edit yet another DNS entry incorrectly.. Thanks

My hosts' DNS manager allows adding a name and a value for a TXT record, I am confused because I need to add a DKIM entry such as this:

mail._domainkey.foo.com. IN TXT "k=rsa; t=y; p=MIGfM...(longrsapublickey)...f323a0

Is the whole thing a "value" or is the mail._domainkey.foo.com. name and after "IN TXT" is the value? I had tried that and it said ignoring out of zone data so that cannot be it..

I created a new user named bob, and gave it its own /home/ folder on my server so I can SSH into that instead of root.

I had added bob into the nogroup group as I had read somewhere (or poorly interpreted it) as allowing me to edit nobody's files, but I get permission denied when I try.

How do I edit nobody's files without running sudo every time, how do I grant permission for bob to edit them?

I just recently deployed a LAMP stack that came as a package for Linode (a VPS host), I can successfully run Apache 2 but the process and all my files within are owned by root. On my shared web host I could run as nobody or similar to help restrict access, how can I do this myself on my VPS?

I think the steps would be:
- Create user apache2
- Write in httpd.conf to use apache2?
- recursively chown all my files to apache2.

I simply can not find a solution except to use another package and have it installed like that, I would really enjoy some guidence on this. Thank you.

If I am mistaken if this is already the case I apologize. I do most of my work on a Linux system and use Filezilla to often upload my work, unfortunately I sometimes open a few files on my Windows PC when that is available and it messes up newlines completely. When I reopen them after being saved in Windows it becomes a garbled mess with no newlines.

Is there a plugin or option for Filezilla or a client you know of that will detect and take care of saving newlines automatically? It would save a lot of worry from having to dos2unix and unix2dos ALL the time.. and forgetting giving me consequences etc.