arikfr's questions

I have a server running UDP server on port 50000. This server receives updates from clients, and pings back the clients every 5 seconds to the socket they were connected from.

Now I want to redirect all the traffic to and from this UDP server via another server. Basically I want to setup a transparent proxy in front of this server. The reason is that I want to be able to easily control which server handles the UDP traffic, without the need to change DNS settings.

At first I tried usingn socat to create this proxy:

socat UDP-LISTEN:50000,fork UDP:myserver:50000

But this seems to work well for incoming traffic, but back traffic from my server to the clients doesn't reach its destination. Also it seems that socat forks another process for each connection, so I might end up with too many processes and also dead processes when clients disconnect/switch IP.

I realized that I can use iptables NAT functionality for this, but it doesn't seem to work for me. The configuration I've tried is:

sudo iptables -t nat -A PREROUTING -p udp --dport 50000 -j DNAT --to-destination myserver:50000

But it doesn't seem to work -- I don't even see the connections on my server. I'm not sure if it's due to wrong configuration or due to some other issues.

Some questions:

1. Do I need to somehow restart iptables once I add the nat configuration?
2. Do I need to allow traffic to the incoming port or having the NAT setting is enough?

UPDATE: The server that actually handles the traffic is on a different machine than the one that I'm trying to setup iptables on.

Assuming my Postfix server serves the domain example.com, I want it to accept mail for all the subdomains of example.com except for mail.example.com and in.example.com.

At first I've tried using a regular expression that excludes these two subdomains:

(?(?!mail|in).+\.example.com)

But it didn't work. I wasn't sure if the problem is with the regular expression or whether Postfix's regular expression engine doesn't support conditionals, so I've decided to try and use conditionals of the table file. At this point my table looked like this:

/etc/postfix/virtual_domains:

if !/(in|mail)\.example\.com/
/.+\.example\.com/ OK
endif

/etc/postfix/main.cf:

virtual_alias_domains = regexp:/etc/postfix/virtual_domains

But it didn't work either. What am I doing wrong?

For monitoring my Postfix install I'm using pflogsumm, which parses the Postfix's logs and generates a nice summary.

I tried to find a similar tool for Dovecot, but couldn't. Is there such a thing?

Some clarifications:

1. The tool has to be open source.
2. We're using Dovecot only for POP3 protocol.

In the technology preview posts about Passenger 3 they discussed zero downtime:

In Phusion Passenger 3 we’ve implemented zero-downtime web server restart. Phusion Passenger and the web server are restarted in the background, and while this is happening, the old web server instance (with the old Phusion Passenger instance) will continue to process requests.

From what I understood, when using Passenger 3 no requests are delayed or dropped during deploy time. But from my tests with Passenger 3 and nginx this is not true. What I'm experiencing is that already made requests are completed, but new requests that are being made when deploying are delayed and waiting for Passenger to spawn new instances.

What I was expecting is that the old Passenger instances will keep serving requests until the new one is fully ready to accept incoming requests.

Did I misunderstand their explanation or I have some bad configuration?

Just to clarify, I understand that I can achieve zero downtime with multiple servers, but in this case I'm referring to one.

Is there a way to limit the number of concurrent connections to a POP3 mailbox on a Dovecot server?

The current situation is that if two clients connect to the same mailbox, then the server accepts both connections, but only one can run commands on the server (like LIST, RETR, etc). If you try to use of one these commands from the second connection, it will be locked until the first one disconnects.

I was wondering if there is a way to make the server refuse the second connection.

Some of the connections between my application server (Ubuntu, Apache, PHP) and my DB server (Ubuntu, MySQL) are failing with error code 4. According to the documentation error code 4 is:

OS error code 4: Interrupted system call

At first I thought that maybe the issue is that the DB server has too many connections and fails because there are too much open files. But it seems not to be the case because:

1. Too many open files has different error code (24).
2. I've checked and during peak time the server had 497 files open (checked using lsof command) while the maximum is 1024.

The TCP settings were already checked (see prior question).

Any ideas what this can be or what should I check?

I would like to know if there is some utility that can show how many open files I have on an Ubuntu server?