Eonil's questions

I am a newbie server manager and trying jails. It's too hard to configure it. So I decided to use ezjail utility. It seems working, but I got an issue.

When I installed my FreeBSD box, it was 9.0-RELEASE. After I upgrade it, it became 9.1-RELEASE-p7.

The problem is ezjail tried to download FreeBSD distro from FreeBSD FTP site. And it failed because ezjail wanted 9.1-RELEASE-p7 and the site has only 9.1-RELEASE.

I tried forcing download by manually specifying 9.1-RELEASE path, but created jail doesn't work with this message.

Configuring jails:.
Starting jails: cannot start jail "example_com":
.
Error: Could not start example.com
  You need to start it by hand.

I have no idea why it doesn't work. What is the recommended way to configure a jail for beginner?

P.S. My test VM is using DHCP and doesn't have static IP and domain name. can it cause this problem?

Update

I succeeded to install a jail using bsdinstall jail $JAIL_NAME. Anyway, I have no idea how it's done...

How can I get list of open SSH tunnels?

I'm using Mac OS X client connected to FreeBSD server. I'm trying to query open tunnels on the client.

I have open tunnel with

ssh -2 -N -L 5001:localhost:1019 [email protected]

The tunnel is working good, but I have to Ctrl+C or Ctrl+D to force quit the tunnel. How can I close the tunnel gracefully?

I just know there is two kind of firewalls. Stateless and stateful. It's hard to determine what kind of firewall I have to use. Currently I have to run firewall within same machine runs services, Basically I want stateless because of its less resource consumption. However if it is not sufficient for security, it's meaningless. I'll run HTTP, SSH, NFS (only over SSH), and some custom made server on several TCP/UDP ports. Should I use stateful firewall?

(edit) Maybe the question can be assumed as 'Should I use stateful rules?'.

How can I make NFS connection secure? Remote server is on the internet, and not in local network. The server has no firewall and connected to internet directly.

I'm studying making a CGI program. Now I can print text what I want. But how can I get GET or POST input from CGI program? (such as query-string or post-body)

freebsd-update command prints updating entries after it downloaded. But it shows them with more. So I have to press some spaces to continue. Is there a way to avoid this behavior? I want to execute update fully automated manner.

How to automate SSH login with password? I'm configuring my test VM, so heavy security is not considered. SSH chosen for acceptable security with minimal configuration.

ex)

echo password | ssh id@server

This doesn't work.

I remember I did this with some tricks somebody guided me, but I can't remember now the trick I used...

I created a new database test and created user 'eonil' with this command:

CREATE ROLE eonil LOGIN ENCRYPTED PASSWORD 'password' NOINHERIT VALID UNTIL 'infinity';

on my PostgreSQL. I run psql -U eonil test. When I tried to make a new schema, it shows an error.

test=> CREATE SCHEMA new_schema AUTHORIZATION eonil;
ERROR:  permission denied for database test
test=> 

Why does this make an error? What's required?

All I know about differences of them is varchar has limit, and text is not. The documentation does not mention about this.

Is that really the only difference? No consideration about performance or etc?

How to print list of users and groups on FreeBSD?

How can I download a single port from freebsd site instead of downloading the whole port collection?

I have a VPS hosting account, and the hosting is offered by Xen, and controllable via VNC. I could connect to my machine with my iOS app, but couldn't with Finder on Mac OS X. Maybe there is an unknown problem on Finder's VNC client. However I need Xen compatible VNC client for Mac OS X. Please recommend some.

What's the most simple way to extract substring on Unix shell (with regex)?

Simple means:

• less feature
• less options
• less study

Update

I realized regex itself is conflicting with simplicity, and I chose the simplest one cut as the chosen answer. I am sorry for vague question. I changed title to represent current state of this QA more precisely.

I have a script:

ACTION="build"

env -i
exec ./makelib.sh Release "${ACTION}"

env -i
exec ./makelib.sh Debug "${ACTION}"

Second exec does not execute. Why, and how can I execute it?

As I know, newly running shell script inherits it's environment variables. Is there a way to block this? (running shell without variable inheriting)

Currently, web application need to offer some kind of cross-domain HTTP header to access data on other domain: http://openfontlibrary.org/wiki/Web_Font_linking_and_Cross-Origin_Resource_Sharing

Is there any way to configure CouchDB to support unlimited cross-domain access? I'm using the db in-house purpose only.

How can I download multiple files stored in a text file with curl and xargs? This is my last trial:

cat listfile.txt | xargs curl -O

first file works well, but other files are just output to stdout.

I'm console newbie. As I know, matching multiple patterns like this:

aaa|bbb

But, | character is pipe on console, so how can I specify those multiple patterns for grep?

I'm using Mac OS X. I'm trying to copying some files with cp command for a build script like this.

cp ./src/*/*.h ./aaa

But this command fires an error if there is no .h file in ./src directory. How to make the command don't fire the error? (silent failure) The error makes build result fail, but I just want to copy when only there are some header file.