When I said in my previous post that the Ubuntu 9.10 upgrade went fine, I spoke too soon…… It seems that the normal Sys-V init system has been replaced by upstart, which is Ubuntu’s attempt at speeding up the boot process by initialising services in parallel. The result being, that /etc/rc.local is no longer the last script to run, and the init scripts are pretty much run in random order, causing all sorts of race conditions.

I finally decided to upgrade my fileserver from Ubuntu 9.04 to 9.10, as 10.04 LTS is due end of April, so its easier to upgrade version-by-version than jump two. It actually went without issue, I had already gathered the .deb’s needed for PIPS and iScan for my printer/scanner and tested them on a 9.10 virtual machine, although by the looks of it, Gutenprint5 now has Epson RX425 drivers built-in, however I’ve seen reports that the colours are too dark.

Had a few problems with Fedora 12, so thought I’d make a note of some fudges fixes I’ve found: Lag issue with Skype – people hear you a couple of seconds after you talk, or you press a number button and don’t hear the sound for a couple of seconds. No problem on F10. Discussed in this forum thread, fixes from here seemed to work for me – or a fudge is to prevent Skype being the first program to use PulseAudio.

Today I’ve been building a noarch RPM specfile for setting up a Nessus environment with custom [signed] plugins. This is quite a good starting point, not really a tutorial though. rpmbuild signs the plugin using nasl as per these instructions, but to automate it I had to remove the passphrase from the private key I was using, as rpmbuild won’t be able to prompt for a passphrase! The command to do that is:

I’ve recently upgraded to nmap 5.21 on my Fedora 12 box, and it turns out that the RPM’s distributed by Fyodor can’t resolve hostnames (even localhost, uname or /etc/hosts entries) apparently due to his build environment being CentOS 5.3 which has a difference glibc and gethostbyname() implementation. A “fix” is to start the nscd service. The problem seems to exist in at least 5.00, 5.20 and 5.21 RPM’s of nmap, however ncat works fine.