Installation freezes

Using Ubuntu EC2 ami-6407f20d (10.04 LTS) freshly made. Set fully qualified hostname, assigned an elastic IP, ran install script:

...
INFO - Removing Debian standard Webmin package, if they exist (because they're broken)...
INFO - Removing Debian apache packages...
INFO - Installing dependencies using command: /usr/bin/apt-get --config-file apt.conf.noninteractive -y --force-yes install postfix postfix-pcre webmin usermin ruby libapache2-mod-ruby libxml-simple-perl libcrypt-ssleay-perl unzip zip quota
...in progress, please wait...
......................................................................

Hangs there.

Went to log (name has changed?) -

ubuntu@samfordwebsites:/var/log$ sudo tail vmbuilder-install.log
2010-09-23 13:02 INFO    : + dist_ge lucid maverick
2010-09-23 13:02 INFO    : + [[ lucid > maverick ]]
2010-09-23 13:02 INFO    : + [[ lucid == \m\a\v\e\r\i\c\k ]]
2010-09-23 13:02 INFO    : + dist_ge lucid maverick
2010-09-23 13:02 INFO    : + [[ lucid > maverick ]]
2010-09-23 13:02 INFO    : + [[ lucid == \m\a\v\e\r\i\c\k ]]
2010-09-23 13:02 INFO    : + '[' -f /tmp/vmbuilderVnuFCK/root/boot/grub/menu.lst ']'
2010-09-23 13:02 DEBUG   : not installing nx (not -desktop)
2010-09-23 13:02 INFO    : Copying to disk images
2010-09-23 13:02 DEBUG   : ['rsync', '-aHA', '/tmp/vmbuilderVnuFCK/root/', '/tmp/vmbuilderVnuFCK/target']

Been there 30 minutes and I don't think its moving.

Status: 
Active

Comments

It may take a while to apt-get install all the packages.

Use top to see what process is using the most CPU time. Also, run ps -ef | grep dpkg to see what is being installed ..

Thanks very much for the prompt reply.

It had broken. When I attempted to shell in to check things I got:

~/aws $ ssh -i /home/paul/aws/ec2-keypair.pem ubuntu@ec2-184-73-108-61.compute-1.amazonaws.com
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
@    WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED!     @
@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
IT IS POSSIBLE THAT SOMEONE IS DOING SOMETHING NASTY!
Someone could be eavesdropping on you right now (man-in-the-middle attack)!
It is also possible that the RSA host key has just been changed.
The fingerprint for the RSA key sent by the remote host is
a7:d8:1a:36:4e:a8:ff:7f:cf:32:82:3a:62:7b:f2:8c.
Please contact your system administrator.
Add correct host key in /home/paul/.ssh/known_hosts to get rid of this message.
Offending key in /home/paul/.ssh/known_hosts:14
RSA host key for ec2-184-73-108-61.compute-1.amazonaws.com has changed and you have requested strict checking.
Host key verification failed.

I tried again with a completely new fresh instance. I just booted it and ran the install script (which is renamed to install_virtualmin.sh - can't imagine that would be a problem).

This time it got a little further:

..............................................................................
INFO - /usr/bin/apt-get --config-file apt.conf.noninteractive -y --force-yes install postfix postfix-pcre webmin usermin ruby libapache2-mod-ruby libxml-simple-perl libcrypt-ssleay-perl unzip zip quota: Succeeded.
INFO -  Succeeded.
INFO - Installing Virtualmin and all related packages now using the command:
INFO - /usr/bin/apt-get --config-file apt.conf.noninteractive -y --force-yes install virtualmin-base
...in progress, please wait...
Extracting templates from packages: 100%
Extracting templates from packages: 100%
....Write failed: Broken pipe

Just to reiterate - this is 10.04 LTS - ami-6407f20d.

I am very keen to get this working and will happily help you if there is anything I can do. It looks like the 'host identification change' is something to do with the problem. On my first attempt I had changed the hostname to reflect the main website domain for the server but I did nothing to the instance at all on the second attempt.

My shell interactions with other EC2 instances where not affected so I don't think this is anything to do with my local system.

Joe's picture
Submitted by Joe on Sat, 10/09/2010 - 02:41 Pro Licensee

That file, "vmbuilder-install.log" is not the virtualmin install log. That's something altogether different and unrelated to Virtualmin. You'd want to find the actual Virtualmin install log to look at, as that one isn't going to have anything useful in it. ;-)

You're getting a new ssh key from the host for some reason, which is what that ssh error is about. That seems like maybe you have a whole fresh instance when you log back in, which is weird. I don't know why the image would be reset. But, again, this error is also completely unrelated to Virtualmin. (There's no way Virtualmin could reset your image.)

The write failed broken pipe bit might be a useful clue, though I don't think I've ever seen it. That stage of the install is slow and can take quite a long time (several minutes anyway), so there may be something timing out, or something along those lines. Though I can't think of what could be timing out or why.

OK. I've used other instances without any problems. I'll try some other things and see if I can move forwards to a point where I can look for the right log (or get it to work).

As I just followed the instructions on a new install it would seem likely that other people have hit this.

Thanks for your help.

OK the bad message was a result of reusing the elastic ip. Fixed by deleting ~/.ssh/known_hosts and starting that over.

Sorry if this is turning into an 'installation help' thread. I am very grateful for your help and I will make any necessary additions or changes to the installation guide for you by way of thanks if you wish:

After several more attempts I've managed to get the install script to complete on a new instance based upon ubuntu's 10.04 LTS EBS ami ami-ea1e60b8.

All the ports are set up as per http://www.virtualmin.com/documentation/aws/virtualmin_gpl_ami. I also added a ping access which works fine. http://175.41.130.167/ brings up the apache default page fine. http://175.41.130.167:10000/ just times out. Turns out the service is not running.

If I start webmin manually it works and I get the login screen although I don't have a user account yet so I cannot actually log in to it.

How do others get past this - is there log I should check or a getting started guide I am missing?

You should be able to login as root using the system's root password - although this being ubuntu, it may not have one set.

In that case, you can set a password for Webmin's root user with the command :

/usr/share/webmin/changepass.pl /etc/webmin root yourpasswordhere

Thanks again:

ubuntu@ws7:~$ sudo /usr/share/webmin/changepass.pl /etc/webmin root nJkL8761 Updated password of Webmin user root Webmin is not running - cannot refresh configuration ubuntu@ws7:~$ sudo /etc/init.d/webmin start

And I'm in! Fantastic.

Why does it not start as part of the script (or at least on a reboot)? I think this should be in the installation guide.

Are there other processes I need to start?

Thanks - I'm moving forwards!

That's odd, as the installer normally does start Webmin automatically.

To make sure everything is OK, SSH in as root and run :

virtualmin check-config

and see what it reports..

It is odd. More on that below but first:

Thank you!!! I am very impressed. By sheer coincidence after I tried Virtualmin on an EC2 server with a view to migrating my existing servers to it - and got it working with your help, a client asked me to configure their VPS on the Intervolve rack in Brisbane to run three of their websites.

Thanks to your help and the practice I used virtualmin and it went very smoothly. So much easier than doing it all by hand.

I am very impressed. The documentation is a bit thin but I like the simplicity and effectiveness of virtualmin/webmin/usermin. Hopefully I will be able to contribute to the documentation over time.

On the oddness: It ran fine at startup on the Intervolve VPS but consistently fails on reboot on the EC2 instance (both are 10.04 LTS and both were installed exactly the same way.

It appears to run fine on the EC2 instance when manually started.

If you want to investigate further I could let you ssh into the EC2 instance so you can see what is going on - its only a test server atm. Email me directly if you wish to do that. Failing that, any hints as to what you think could be happening will be welcome. I will look into it further myself but other work has to take priority over the next few days.

Yes, a remote login to see why it isn't starting at boot would be useful.

Assuming that I can reboot the system for testing purposes..