Toby's questions

I am running some scripts as part of pre/post deploy using DeployHQ, I need to be able to call the script on one line and for there to be no feedback from the terminal so I want to pass the password into the sudo command (instead of sudo asking for the password at the time of use)

I am trying to get sudo to work with stdin by doing something like;

echo my_password | sudo -S chown -R my_user:www-data /srv/www/my_site/public_html/

But what is getting returned is;

[sudo] password for my_user: 
 Sorry, try again.
[sudo] password for my_user: 
 sudo: 1 incorrect password attempt

The password in question is correct and when the command is ran without -S and the user is prompted for the password is works.

Could someone point out what I am doing wrong?

Just to clarify - I want there to still be a password, just one that I can pass in.

I am trying to install make - normally I just type sudo apt-get install make and it works, but on this server I am getting the following;

The following packages have unmet dependencies:
git-core: Depends: libdigest-sha1-perl but it is not going to be installed
          Depends: rcs but it is not going to be installed

Firstly, I don't fully understand what git-core has to do with anything, but for now I just want make to work so I ploughed on by running apt-get -f install but this fails for the same reason.

So I tried to install libdigest-sha1-perl, it fails because rcs is needed. So I go to install rcs and guess why it fails? Yup libdigest-sha1-perl is needed.

I have never come across this before and any pointers on how to achieve this would be greatly received.

The machine is running Debian 5.0.9 if that makes any odds.

We have a small development team that have recently moved to using github to host our projects.

The team consists of three developers, 2 on Windows and 1 on Mac.

I am currently researching the best way to deploy applications to our Linux servers (dev and production).

Capistrano running locally would be ideal but from what I read this won't work for Windows machines.

It looks like the best way is to use a post-receive hook in github, I can see how this would work for auto deploying to dev, but I don't see how we could then deploy to live.

I have found paid projects like http://www.deployhq.com/ but it feels like something that a quick bit of code should be able to do for free, I just can't seem to get myself pointed in the right direction!

I was wondering what would be considered best practice for small team deployment involving multiple local OS's and github.

This is the dumbest thing because I have done this a load of times before and never had issue so I am sure I am just doing something daft but I can't for the life of me figure it out.

I am trying to connect from a Windows 7 machine to a Linux box using key based authentication.

Here are the steps I have followed;

1. Ran ssh-keygen -t rsa from Git Bash.
2. Selected the default area for the file to be stored.
3. Did not pick a password (left it blank)
4. Copied the contents of id_rsa.pub onto the remote server (/home/user/.ssh/authorized_keys)

When I try and log in using ssh [email protected] -v I can see that it tries to match up my keys from the correct location on the Windows box, but it does not accept them and ends up falling back to passwords as the authenitcation method.

I have checked the permissions of authorized_keys - it is set to rw for the owner only.

Any help would be greatly appreciated.

Is there a way to work out if PHP was installed as part of a generic apt-get package, compiled from source or installed some other way on Ubuntu?

I need to recompile it with some more options and I want to try and follow the correct upgrade path.

I am attempting to install both an SSL certificate and an intermediate certificate on an Apache web server. To do this I have had to install mod_ssl (using a2enmod ssl)

First I set that up and restarted Apache, it loaded fine.

Then I set up the certificates and attempted a restart - this appeared in the error logs (error.log);

[Mon Oct 24 04:17:48 2011] [error] Init: Unable to read server certificate from file /etc/apache2/ssl.key/ssl-cert.crt

After some digging I noticed the SSL cert was malformed (everything was on one line), I re-copied everything across for both the main crt and intermediate crt and attempted a restart again, again it failed but this time nothing was written to the error log.

When I run openssl x509 -noout -modulus -in /etc/apache2/ssl.key/ssl-cert.crt I get a modulus returned (something that wasn't happening before I fixed the formatting of text in the files).

If I take out the <VirtualHost *:443> block in my Apache config file, it starts fine.

Apache is running as www-data

Any help would be greatly appreciated.

Linux Info

cat proc /etc/*-release
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=10.04
DISTRIB_CODENAME=lucid
DISTRIB_DESCRIPTION="Ubuntu 10.04.3 LTS"

Apache Info

apache2 -v
Server version: Apache/2.2.14 (Ubuntu)
Server built:   Sep  1 2011 09:45:41

SSL Info

-rwxrwxr-x  1 root root 1.8K 2011-10-24 04:19 ssl-cert.crt
-rwxrwxr-x  1 root root 1.6K 2011-10-24 04:19 intermediate.crt
-rwxrwxr-x  1 root root 1.7K 2011-10-10 09:24 ssl-cert.key

Site Info

<VirtualHost *:443>
    ServerName www.mysite.co.uk
    DocumentRoot /srv/www/mysite/public_html
    ErrorLog /srv/www/mysite/logs/error.log
    CustomLog /srv/www/mysite/logs/access.log combined
    SSLEngine on
    SSLCertificateFile /etc/apache2/ssl.key/ssl-cert.crt
    SSLCertificateKeyFile /etc/apache2/ssl.key/ssl-cert.key
    SSLCertificateChainFile /etc/apache2/ssl.key/intermediate.crt
</VirtualHost>

The first time you try and sudo on a server you get the sudo warning;

We trust you have received the usual lecture from the local System Administrator. It usually boils down to these three things:

#1) Respect the privacy of others.

#2) Think before you type.

#3) With great power comes great responsibility.

Does anyone know the file you need to edit to amend this warning?

If I try and create a link without sudo in /etc/apache2/sites-enabled it fails with a permission denied.

When I try and do it with sudo a file is created but it is not a link.

Any ideas?

The exact command I am running from sites-enabled is ln ../sites-available/newsite newsite

The file I am linking to is owned by me.

I have recently set up vsftpd and whilst everything appears to be running fine and the server is listening on the correct port I cannot connect in using FTP clients.

A connection is established but whilst waiting for a welcome message it times out.

Googling the issue lead to lots of pages saying I should check my iptables, I know little about these but I have printed the results below.

My first question is I am guessing the iptables acts like a software firewall so presumably if I don't have it set up then all traffic should come in fine?

My second question is are iptables the most common but not only way of limiting what traffic can come in, and if it isn't the only is there anywhere else I should be looking?

Please request any additional information and I will edit the question appropriately.

EDIT

I have checked for any related logs and can't find anything of note.

When I try the command ftp localhost I get a timeout, so it looks like it may have nothing to do with IPTables but I don't know enough to totally rule it out.

The results of running telnet localhost ftp and telnet localhost ftp-data were timeouts, but they did try connecting to 127.0.0.1 which is what I would expect.

Here are my IP Tables;

Chain INPUT (policy DROP)
target     prot opt source               destination
fail2ban-ssh  tcp  --  anywhere             anywhere            multiport dports ssh
ACCEPT     tcp  --  anywhere             anywhere            multiport dports ssh,smtp,domain,www,https,ssmtp,xmpp-client,xmpp-server,5280,8999:9003
ACCEPT     udp  --  anywhere             anywhere            multiport dports domain
ACCEPT     all  --  anywhere             anywhere            state RELATED,ESTABLISHED
ACCEPT     tcp  --  anywhere             anywhere            multiport dports http-alt
ACCEPT     tcp  --  anywhere             anywhere            tcp dpt:ftp state NEW
ACCEPT     all  --  anywhere             anywhere            state RELATED,ESTABLISHED

Chain FORWARD (policy DROP)
target     prot opt source               destination

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination
ACCEPT     all  --  anywhere             anywhere            state RELATED,ESTABLISHED
ACCEPT     all  --  anywhere             anywhere            state RELATED,ESTABLISHED

Chain fail2ban-ssh (1 references)
target     prot opt source               destination
RETURN     all  --  anywhere             anywhere

I was wondering what are the common things you should look at when debugging a new subdomain not being picked up correctly by apache.

Here are the steps I followed to create the sub-domain;

I copied the setup file from an existing (working) domain in sites-available.

sudo cp old new

Then I edited this to work with the new site (replace 'old' with 'new')

< VirtualHost *:80>

ServerName new.srv.co.uk

DocumentRoot /srv/www/new/public_html/

ErrorLog /srv/www/new/logs/error.log

CustomLog /srv/www/new/logs/access.log combined

< /VirtualHost>

I then created the link in the sites-enabled folder.

ls -s ../sites-available/new new

The next step was to create the appropriate directories in /srv/www/ and gave them the same permissions as other site already listed there.

Finally I restarted apache

sudo /etc/init.d/apache2 restart

It restarts fine but the URL will not work for the site.

If I edit the file in sites-available to break it (remove the '*' for example) apache fails to restart, so I know it is finding that file OK.

There is nothing in the error log and browsers are reporting things like 'Firefox can't find the server at ...'

I have installed vsftpd and configured as per several tutorials (just the very basic, standard setup) but when I run cat /etc/services there is nothing appearing for Port 21.

What steps do I need to follow in order to add this in?

I have hit an issue on a MS SBS machine where every morning the datastore for the exchange mailboxes dismounts itself.

We believe the issue is that it has grown too large over time and needs cut down a bit.

As part of this we have removed (purged) some mail files that were no longer needed, which should have given us a saving of roughly 3GB (more than enough saving for what we need).

So I deleted the mailboxes, then purged them and noticed that the .edb file was still reporting the same size, I dismounted and remounted it to see if that would have any effect but it did not.

Am I missing a step? I have read online that you can run offline defrag on the file but that seems to only save you a small amount of whitespace.

Any help would be greatly appreciated.

Recently I had an issue where a virtual machine had ran out of space and was suspended, a restart was able to to fix the issue but when I SSH'd in I noticed there were two large files (both delta.vmdk files) in datastore2/vm_in_question.

When I look at the snapshot manager in the VMware Console I see one snapshot from yesterday (Our VM backup tool created this one in order to perform a weekly backup which oddly enough is still running, this is another issue!) but this was a leaf node under two Consolidate Helper - 0 files.

I have never seen these before and was wondering if one of them at least was causing one of the large delta files?

Any information on how these files are created and what I can do to safely remove them and stop this from happening again would be greatly received.

I connect to one linux machine with a standard user account and use sudo when it is necessary to perform something a little above my users station.

One of those things would be to perform a pull using git (sudo git pull).

I want to set up key based authentication between the machine I am in and the git repository so as to not need to put in a password each time.

I have been able to set up authentication between my standard account on the machine and the git repository but when I try and set it up as root it will not work (I perform the same tasks as I did for my normal account but using sudo each time)

I presumed that as sudo runs as root (sudo whoami returns 'root') generating keys and transferring them to the other machine as sudo should work the same as if I was logged in as root. Am I wrong in my assumption or are there issues with using key based authentication as the root user onto another machine?

I am having an issue with a machine on my company network where it has joined the domain OK and the machine itself can happily connect with play with other machines on the network but nothing can connect to it (or indeed ping it).

What I would like to achieve to to be able to remote desktop onto it (obviously I have turned this option on in the settings)

The machine is running Windows XP, the firewall is off and it is getting an IP on our network. I am a bit of a networking n00b so I don't know what other information is relevant.

Any ideas on what could be causing this?

I have several small to medium sized mysql databases, say about 40 and I need to migrate them from one whm server with different cpanel accounts to another one with the accounts set up from the previous server and old versions of the databases I am moving already in place.

Could anyone recommend the quickest way of doing this, I was going to manually dump each one and import it across but it seems very time consuming, I would like to cut out my machine as the middle man if possible and automate were I can.

I was wondering what is considered best practice when several developers/administrators require access to the same web server.

Should there be one non-root user with a secure username and password unqiue to the web server which everyone logs in as or should there be a username for each person.

I am leaning towards a username for each person to aid in logging etc however then does the same user keep the same credentials over several servers, or should at least their password change depending on the server they are on?

Should any non-root user of the system be added to the sudoers file or is it best practice to leave everyone off it and only let root perform certain tasks?

Any help would be greatly appreciated.