[Gllug] Modern Fault finding techniques

[James Courtier-Dutton]

On 17 November 2011 16:50, James Hawtin <oolon at ankh.org> wrote:
> I rescently went for an interview for Linux SA job which i failed to get,
> after asking for feedback aside from saying I did have experience in
> technologies I did not claim to have experence in on my CV (java application
> servers, wish people would read CVs), I got the following.
>
> "He tends to go for traditional methods in solving problems and has no
> concept of newer/ better technologies"
>
> So the question is, what do people consider newer better technologies. The
> interview question was "A user phones up and say the page on the web site is
> very slow".
>
If they did not read your CV, I would not put much weight on their feedback.
It sounds to me that they were looking for a java applications/SOA
hosting service type system administrator.
I.e. Applications hosting support.
It has very little to do with a traditional Linux System Administrator's job.
If "A user phones up and says the page on the web site is very slow"
you can be pretty sure there is nothing wrong with the Linux OS. It is
more to do with a problem with the way the application is
written/designed.
If I was answering that question, I would start looking at the
application metrics monitoring tools, and from that you should get an
idea as to what part of the system is not performing. It might be
database transactions running slowly, or cache lookups not working as
expected.
It would probably be hosted on a cluster of Virtual Machines, and you
could get metrics from that to see if a single VM is misbehaving or
not.
You could also look at the SAN, and see if a HD has failed and if it
was degraded and busy re-imaging to a new disk.
All these possible causes of a "web site being very slow" have nothing
to do with Linux or even GNU/Linux.

Kind Regards

James
--
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug