Tiddo's questions

I am considering moving from apache to nginx. However, there is one problem I don't really know how to handle best: On my server I have some 3rd party software, e.g. phpbb, phpmyadmin etc., and these applications use .htaccess files, which aren't by nginx.

Of course I could decide to change all htaccess rules and files to their nginx equivalent, however doing so also means that I need to manually check and update these rules when I update one of these applications.

Are there any other solutions to this problem?

I'm using htaccess to rewrite my urls so that I can have nice clean urls. However, the same htaccess file does different things on my local server and my remote server:

On my local server the url to the website is like http://localhost/example/ and on my remote server the url is http://example.com/.

For my local server I can use the following htaccess redirect rule:

RewriteRule ^(.*)$ index.php?page=$1 [L,QSA]

However, when I use this on my remote server I get an internal server error. Instead I have to use this: (note the leading slash)

RewriteRule ^(.*)$ /index.php?page=$1 [L,QSA]

Unfortunately this doesn't work on my local server: this rewrite rule requests http://localhost/index.php instead of http://localhost/example/index.php on my local server.

How can I make this work on both my remote and local server?

As a webdeveloper I have, like many, my own VPS. However, I am a webdeveloper, not a system manager. For me it's quite challenging to install a VPS and to keep it up to date and secure. It takes me days, if not weeks, to get a server up and running. I also hardly dare to update my server since I'm afraid I screw things up.

So my question to you is what resources can you recommend me that won't only tell me how to install a LAMP server, but also explains how to keep everything up to date and secure?

edit: Installing everything isn't the main problem, the main problem is configuring everything such that is is secure and that it works. Most tutorials just tell me how to install everything, but usually that's just not enough.

In my specific case I have a CentOS VPS for which I need LAMP, ftp and email. I also host multiple domains on one VPS.

Possible Duplicate:
Wildcard DNS and subdomains

I'm trying to create a domain alias, but I don't really know how to do this. I want all trafic on all ports and all subdomains to be redirected to another domain at the same port with the same subdomain (so any xxx.domain1.com:yy becomes xxx.domain2.com:yy).

Is this possible, and what is the best way to do this? I have full dns and root access to my server, if that's needed.

I've set up a server with postfix for sending mails. However, when I try to send an email to a gmail account, I get the following errors:

May 16 23:41:47 vps4195 postfix/smtp[15949]: certificate verification failed for gmail-smtp-in.l.google.com: num=20:unable to get local issuer certificate
May 16 23:41:47 vps4195 postfix/smtp[15949]: certificate verification failed for gmail-smtp-in.l.google.com: num=27:certificate not trusted

After these errors the mail is delivered to the gmail accounts, but the mail will almost always go to the spam directory, even when no mail filter is used inside gmail.

So how can I get rid of these errors?

ps. Probably this isn't enough information for you to help me fix this, but I'm not really an expert on this subject, so please tell me if I need to post more information.

edit: The system postfix is running on is CentOS 5.5 (VPS)

I have a small vps, with only 256mb of ram, with maximum burst up to 512mb. When I configure my vps without innodb, it only uses 130 mb of ram, so that is no problem for me. But when I turn on innodb, The memory usage grows to about 300-400 mb. Is it possible to run innodb such that I won't exceed the 256mb? preferably I don't want to use more than 100mb for innodb. I already came across some sites which said I could limit the memory usage, but if I limit it to only 100mb will the db run well enough? (compared to for example the MyISAM storage engine)

If 100mb is too little memory for innodb, can you recommend me any other storage engine which supports transactions?

I am configuring mysql for my server, but I have one question: Is it possible for each domain to see their own mysql server, but only having to run one service at the server? I want to keep all databases for the domains separated, but I don't want to have to prefix each database name.

eg: suppose I have domain1.com and a domain2.com running on the same server. For both domains I want a database called statistics. If I want to do this now, I'll have to prefix the databases, so there are two databases: domain1_statistics and domain2_statistics, and the domains can only access their own databases. But I want to create 2 databases called statistics, one for each domain, so that each domain only sees its own statistics database. Is this possible?

I currently have a server with postfix/dovecot/postfixadmin and squirrelmail running for multiple domains. I use mysql to store the account/domain details. However, if I go to the webmail of a domain on my server, I can login to all email accounts on any domain on my server. How can I make it only possible to login to a mailaccount using on the right domain? And if i've done this, is it also possible to allow to login to all mailaccounts using the ip-address of the server directly?

In short, I want that if a user goes to domain1.com/webmail he can only login to emailaccounts with @domain1.com, but if a user goes to A.B.C.D/webmail he can login to all emailaccounts on the server.

edit:
I think I wasn't clear enough in my post, so I will try to explain a little bit better what I have:
I have one postfix/squirrelmail installation, and squirrelmail can be accessed on any domain through domain.com/webmail (/webmail is aliased). But because I only have one installation, you can login on all domains on my server using any domain. So for example, if I have domain1.com and domain2.com, you can login with [email protected] at www.domain2.com/webmail. I dont want this. The only exception to this rule is that I want to be able to access the complete mailserver, so all emailadressess on all domains, using the ip-address of the server (A.B.C.D/webmail). Sorry if my question isn't clear, english is not my main language so I sometimes have some trouble explaining my problems in english.

I am configuring postfix and dovecot on my server, but for some reason I can't receive mails. I've followed these instructions for configuring the mailserver. I've tested with pingability.com that port 25 is open (didn't get any errors for the mail), and I am able to send mails. I just can't recieve them, I don't see them in the maildir and I don't see them on squirrelmail.

I am not really a linux expert, so can anyone help me to resolve this problem?

My OS is CentOS 5.5

edit1: According to the logfile /var/log/maillog the mail is received, but probably not deliverd. If I check the logfile directly after i've sent an email, the last lines of the logfile look like this:

Mar 27 23:59:24 vps4195 postfix/smtpd[15822]: connect from mailserver[ip]
Mar 27 23:59:25 vps4195 postfix/smtpd[15822]: 01EE0AD18170: client=server[ip]
Mar 27 23:59:25 vps4195 postfix/cleanup[15826]: 01EE0AD18170: message-id=<someid>
Mar 27 23:59:25 vps4195 postfix/qmgr[7544]: 01EE0AD18170: from=<emailaddress>, size=1980, nrcpt=1 (queue active)
Mar 27 23:59:55 vps4195 postfix/smtpd[15822]: disconnect from server[ip]

edit2: The logfile also contains many of these entries, but I don't really know what to do with them:

Mar 28 00:00:16 vps4195 postfix/virtual[17954]: fatal: bad string length 0 < 1: virtual_mailbox_base =
Mar 28 00:00:17 vps4195 postfix/master[7538]: warning: process /usr/libexec/postfix/virtual pid 17954 exit status 1
Mar 28 00:00:17 vps4195 postfix/master[7538]: warning: /usr/libexec/postfix/virtual: bad command startup -- throttling

edit3: I've fixed the previous error (added a virtual_mailbox_base to the main.cf file), but now the log file gives me this error when it tries to receive an email:

Mar 28 00:11:33 vps4195 postfix/virtual[3089]: E8025AD180BB: to=<email>, relay=virtual, delay=0, delays=0/0/0/0, dsn=5.1.1, status=bounced (unknown user: "email")
Mar 28 00:11:33 vps4195 postfix/qmgr[3077]: E8025AD180BB: removed

edit4: got a little further this time, it looks like I missed a line in the main.cf, but then I got this error:

Mar 28 00:36:15 vps4195 postfix/qmgr[30545]: D7E6DAD18077: from=<mail>, size=755, nrcpt=1 (queue active)
Mar 28 00:36:15 vps4195 postfix/virtual[30610]: warning: recipient mail: not found in virtual_uid_maps
Mar 28 00:36:15 vps4195 postfix/virtual[30610]: D7E6DAD18077: to=<maill>, relay=virtual, delay=0.07, delays=0.05/0/0/0.01, dsn=4.3.5, status=deferred (mail system configuration error)

So I created the virtual_uid_maps and virtual_gid_maps entries in the main.cf file, and everything works now!

I am using VirtualBox on windows 7 64-bit, with CentOS5.5 as the guest os. I try to setup a bridged networking connection, but when I select bridged adapter, I can't choose any networking adapter in the second list. It just says not selected. What am I doing wrong?

I have virtual box installed on a windows 7 64-bit computer, with Cent OS 5.5 as guest os. I want to be able to use ftp between those. I've installed vsftpd on the guest os, and the guest os uses a nat connection with the host os for internet. So far, I am able to connect to the guest os using ftp (in filezilla), but after the list command is executed, nothing happens, until the command is timed out. This happens in both active and passive mode.

I do have set a pasv_min/max_port in the vsftpd.conf file, listing is enabled, and the ports are redirected in virtualbox. Also the ftp_data_port is set to 20. I also tried setting the pasv_address, but I had to set it to 127.0.0.1, but than filezilla gives me this:

Command:    PASV
Response:   500 OOPS: bad family
Command:    PORT 127,0,0,1,139,204
Response:   500 OOPS: child died

Can someone help me with this?

I'm trying to understand DNS a bit better, but I still don't get A and NS records completely.

As far as I understood, the A record tells which IP-address belongs to a (sub) domain, so far it was still clear to me. But as I understood, the NS record tells which nameserver points belongs to a (sub) domain, and that nameserver should tell which IP-address belongs to a (sub) domain. But that was already specified in the A record in the same DNS file. So can someone explain to me what the NS records and nameservers exactly do, because probably I understood something wrong.

edit: As I understand you correctly, a NS record tells you were to find the DNS server with the A record for a certain domain, and the A record tells you which ip-address belongs to a domain. But what is the use of putting an A and an NS record in the same DNS file? If there is already an A record for a certain domain, then why do you need to point to another DNS server, which would probably give you the same information?