Posts Tagged ‘oracle’

Opennebula, ZFS and Xen – Part 3 (Oracle VM server)

September 27, 2010

Read OpenNebula, ZFS and Xen, Part 1
Read OpenNebula, ZFS and Xen, Part 2

Oracle VM server is Oracle’s virtualization platform. As with most Oracle Linux offerings it’s essentially a Red Hat Enterprise Linux, which however bundles a more recent version of Xen.

# more /etc/enterprise-release
Oracle VM server release 2.2.1
# /usr/sbin/xm info | egrep 'major|minor|extra'
xen_major              : 3
xen_minor              : 4
xen_extra              : .0


Installing the Oracle VM server is pretty straightforward. Download Oracle VM server, burn the ISO image to a CD/DVD and boot the hypervisor host from it. The setup is pretty straightforward if you’ve installed RHEL or CentOS before, the only catch being that you don’t really get to select any packages to install.

Integrating with Opennebula

Integrating Oracle VM server with our Opennebula setup is similarly simple.

  • Add Oracle’s public repo to YUM
  • Use yum to install ruby
  • Remove requiretty from the sudo configuration
  • Add an entry for the Shared FS to /etc/fstab
  • # grep cloud /etc/fstab
    storage-server:/export/home/cloud /srv/cloud       nfs4    auto          0 0
  • Tweak rcpidmapd.conf so that it matches the NFSv4 domain of the storage server
  • Create the oneadmin user and cloud group

Farewell old friend

January 29, 2010

James Gosling piece touched a sensitive chord; finding out that is no more touched another.

While my first contact with Unix was also through a VT220 terminal connected to a SunOS server it was not love at first sight. Even when I ended up loving a distant descedant of Unix I recall myself poking our Solaris admin at what used to be dayjob 6 years ago over the inefficiency of the Solaris userland.

It was not till 2005 that I started dealing more frequently with Solaris, version 8 back at the time [1]. And then Solaris 10 came around. x86 became a first class citizen, allowing for a huge performance boost in afforable hardware. With it slowly came Zones (a superior virtualization technology reminiscent of FreeBSD jails), DTrace, Grub (allowing peaceful co-existence with other O/S), ZFS (literally the last word in filesystems, at least in the English alphabet ;)) and more. All of these coupled with typically superior documentation and always sticking to the POLA for long-time users. And the same way I ended up hating Windows and loving Linux, in spite of starting my systems administrator carreer in a Windows environment, I ended up loving Solaris. Not that I hate Linux nowadays [*], I’ve just grown too old to accept things breaking or changing for little reason every now and then.

In every end lies a new beginning they say. Let’s hope that Elisson manages to monetize the numerous cool technologies that have been coming out of Santa Clara but at the same time the spirit of quality, technical and design excellence, the well-thought out customer support and the ever-present drive to push our overall computing experience to new frontiers that “Sun” represented will stay with us.

[1] Solaris 9 is like the new Star Wars trilogy; passionate Solaris users -such as the author- vehemently deny it ever existed.

[*] OK, that’s not entirely true; I do hate one particular “flavor” of it.

On monty’s petition to save Mysql

January 2, 2010

I have grown so tired of reading about this that I decided a short blog post is in order.

Recently Dimitris, a good friend and close associate, posted to a popular Greek Open Source list a link to Monty’s plea for saving MySQL [1].

I was a little reluctant to answer at first, since it’s well known that I am a fan of everything under the Sun and could be accused of being biased [2]. That said, George Keramidas‘ wrote a rather insightful comment:

My personal opinion is that Monty is fear mongering, not because he truly believes that he is saving MySQL but for an own personal agenda.

This gave me a little bit of courage, seeing that there are others who think like me. So I posted a short reply:

It’s actually quite simple. With Sun, a company in a really sad state (financially) Monty could make some money out of MariaDB. With Oracle he just doesn’t any chance.

Besides, claiming that “GPL and dual licensing were good when I was making shitloads of money but now that someone else does it’s suddenly bad” is hypocritical, to say the least”

The follow up by Dimitris kicked off with a rather surprising comment:
There may be some truth to the above statement (actually, it’s obviously true) [editor’s note: he was referring to the “could make some money” paragraph]
Afterwards Dimitris tried to make a case that in spite of Monty’s personal agenda signing the petition could make for a better future of MySQL:
Things aren’t black and white and this fight is not just for the code but for MySQL as a project (code, community, trademark, servers, foo). There are some examples of projects that were successfully forked with the right leadership but there are many more that remained stale.
There is some merit to the above statement. Yet, I still don’t think that signing the MySQL petition will help at all. As any other FOSS project MySQL can and will thrive if either (or both) of the following conditions stand true:
  1. It gets the backing of a large company, like Oracle
  2. It successfully forms a large community

Signing the Help MySQL petition certainly doesn’t help towards the first direction and is irrelevant to the second (there are countless of successful FOSS projects with either a GPL or a BSD-style license). It only helps:

Now, if anyone is really interested in helping the above interests, feel free to sign the petition. If you ask me, you won’t be cutting MySQL any favors.

P.S. Dimitris and George, I liberally translated your e-mails. If you feel that a translation is off, feel free to let me know and I will correct it promptly.


[2] Disclaimer: other than being a fan of numerous technologies they’ve brought us during the last 25 years, I have no affiliation with Sun Microsystems