[Wylug-help] server distro's - was grub problems...

[Simon Greenwood]

On 12 July 2012 12:06, Richard Ibbotson <richard.ibbotson at gmail.com> wrote:

> > On 11/07/12 22:19, Gary Stainburn wrote:
> > > I've always use RedHat followed by Fedora for my servers (and
> > > workstations for that matter). I did try moving to CentOS but was
> > > frustrated by the restrictive RPM base. I haven't tried anything
> > > else yet, but was wondering what people's thoughts/suggestions
> > > were
>
> On Thursday 12 July 2012 11:28:47 Chris Davies wrote:
> > The reason I moved to Debian, and similar reasons have been quoted
> > by the SysAdmin team, is that the package management "just works",
> > and hides much of the dependency hell from the administrator. I
> > believe yum now does a fair job on RH-type systems but I for one
> > got absolutely fed up with the number of rpms that needed
> > downloaded individually each time a new package was required. Once
> > bitten, and all that.
>
> My own 10p's-worth on this is that I started on RedHat and Debian back
> in 1995.  At that time the RPM package manager was a pile of *******
> ****.  To be nice about it. Present day I find it hard to choose
> between apt-get and yum.  Sometimes I use smart.  TBH I find that
> Fedora or RedHat is easy to use for servers.  Debian is also good.
>
> Last week I tried to install a Debian Apache server with a Debian 6.05
> netinstall image.  It produced a non-bootable hung system.  Had to use
> System Rescue CD to get it to boot.  Probably a good idea to remember
> that this is x86 hardware.  Which probably isn't seen very much in
> commercial circles.  Outside of a University.  Okay, so Universities
> and organisations like NASA use x86 PC clusters but probably not in
> commercial organisations.  I worked in a BT data centre in Sheffield a
> few times.  Not much x86 hardware in the servers that are being used.
>
>
You say that, but BT had a project to consolidate their internal systems a
few years ago and while it started on Sun kit, it ended on HP x86 blades. I
won't say it was entirely successful as it wasn't but there's been a trend
in enterprises to get away from midrange to x86 clusters or blades, mostly
driven by HP and organisations who got stung by Sun's changes in licensing
after the Oracle takeover. Again, it's not always successful but if I was
building that sort of thing these days I'd be inclined to use CentOS or
RedHat on carrier grade x86 hardware for the price.

s/

-- 
Twitter: @sfgreenwood
"TBA are particularly glib"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.lug.org.uk/pipermail/wylug-help/attachments/20120712/1659607f/attachment.htm>