MalphasWats's questions

I have an ancient web application that a few people still rely on. A year or so ago I discovered that the database that backs it had filled almost all of the disk space available. I deleted a lot of the no longer needed content from the database and that was when I learnt that MySQL doesn't deallocate disk space.

I was able to free up about 14Mb which was enough for Apache to write its log files.

It has limped along since and the users that have been using it have stubbornly continued to despite my gentle suggestion that they use the new thing I built that works a bit better!

Today something has eaten the magic 14Mb and everything has stopped working.

Is there anything I can do to reclaim some of the space from MySQL without killing all the data in it?

UPDATED with some of the details asked for in comments:

du -hs /var/lib/mysql : 30G

df -h : size 36G, used 34G, avail 0, use 100%

SHOW VARIABLES LIKE 'version': 5.5.28-0ubuntu0.12.04.3

I've tried a few times to get postgresql running on a couple of different linux boxes, but I've always ended up binning it and going back to MySQL because I can't get my head around how to create and manage users.

This time around, I've actually gotten as far as being able to log in with my own account, but I created the account with the createuser command and it has a different password to my actual user account.

What I'd like to be able to do is log in as me to create databases, users and assign database permissions, but then have an account that is separate and doesn't have an account on the server to use with my web app.

Does anyone have a good step-by-step guide for complete morons that goes through creating users and assigning access to various databases in postgresql?

I have a headless Ubuntu server running on my network. Initially I had configured it for SAMBA for use with my windows machines and that seemed to work ok. I've just bought a Mac and am trying to access the server - it appears in Finder without me doing anything but when I log in, I can't see any files or write to it or anything.

I tried running through this tutorial to set up netatalk and avahi but it fails at the first step - I can't get the source files because I get an error:

Package libcupsys2-dev is a virtual package provided by:
  libcups2-dev 1.4.2-8
You should explicitly select one to install.
E: Package libcupsys2-dev has no installation candidate
E: Failed to satisfy Build-Depends dependency for netatalk: libcupsys2-dev

and I don't know what to do next. I tried going through the steps just using the basic version from the repos which gets me as far as the server appearing in Finder, but when I try to log on, I just get an invalid username/password error (and I've tried so many times I am certain I have the right ones!)

Does anyone have any tips for setting up an Ubuntu server so I can use it with my Mac?

For the Bounty: I'd like the server broadcast via Bonjour so I can see it in finder, and I'd like the option to set it up as a TimeMachine backup if possible.

Thanks

I have an issue with my Apache config, specifically how it interacts with the SVN DAV mod.

What I want to be able to do is have domain.com point at my website, svn.domain.com to point at my svn directory (which is currently outside of my public_html directory) and for both to be accessible via https.

I've got it working for the most part, but if I set up an explicit VirtualHost for the svn subdomain, it stops the svn application from being able to access the repositories (instead just getting a "repository moved permanently" error), I can browse the repos ok.

my config looks something like:

<VirtualHost *:443>
    ServerName domain.com
    ServerAlias www.domain.com svn.domain.com
    DocumentRoot /path/to/website/dir
    ...
    #stuff for SSL cert etc.
</VirtualHost>

<VirtualHost *:80>
    ServerName domain.com
    ServerAlias www.domain.com
    DocumentRoot /path/to/website/dir
    ...
</VirtualHost>

<Location /repo>
    DAV svn
    SVNPath /path/to/repos
    ... #auth details
</Location>

This works mostly, but browsing to svn.domain.com brings up my website, and browsing to domain.com/repo brings up the repository (svn.domain.com/repo works fine, domain.com also works fine), potentially I'd maybe want domain.com/repo to display a set of webpages all about my project, docs etc, svn.domain.com would ideally point to either nothing, or a list of the repositories available or a separate page or something. I thought I'd got the hang of apache config files, but this one is stumping me!

I've set up a repository to configure my websites, the repository is /subversion/web and at the moment I have just one website controlled in /subversion/web/mywebsite

I've set up a post-commit hook script that exports the site into a web directory on the server for testing before I publish it all separately and it works fine.

What I'd like to do is set the script so that I can control a second website in the same repository but only have 'mywebsite' export when I check in a file related to that project.

I've set it up this way because I access the repository from a number of computers via http, it makes it easier to create new projects without having to create a whole new repository and configure apache to point at it and all the rest of it.

latest svn running on latest ubuntu with latest apache2

I'm trying to set up my Win7 work laptop so I can access my svn repos. My repository is set up on a Linux box at home and works fine from all of the my machines, including my work laptop when I'm not connected to the work network.

The problem comes when I try to access through the proxy - I've added the proxy settings in the TortoiseSVN settings panel, when I try to check out a project, it starts downloading and runs for about 6.5MB before giving up and complaining that the remote connection was closed. The repository I'm testing with only has a single text file in it, it's nowhere near 6MB. From what I can work out in my router logs, the requests from work weren't getting through, so I don't know what it was downloading!

It appears there is an issue with the web proxy at work, I'm just wondering if there is anything I can do about it? It's a Microsoft ISA server, and I don't have much control over its configuration, so I'm really hoping there is a solution, I'm just not sure where to start.

Edit: I can browse the repository via HTTP through my work proxy just fine, just can't seem to do anything using svn. I also tried installing SVN proper and doing it from a DOS command line, that didn't work at all, I got the remote connection closed error after just a few seconds and no downloading of phantom giant files.

Thanks