Stuart Bishop wrote:
-- Start of PGP signed section.
> Arnaud Lesauvage wrote:
> > Hi list !
> >
> > I have a small enterprise network (~15 workstations, 1 server), all
> > running windows OSes. Most of our work is done on a PostgreSQL DB (on
> > the windows server).
> > I am the only IT here, and my boss asked me to find a way to have the
> > database always online, without my intervention.
> > Last time I went on vacation, the server crashed and no one was able to
> > repair it.
>
> If your application is normally reliable, I think the best, cheapest and
> simplest way of keeping the system online when you are on leave is to give
> your work the phone number of a company offering PostgreSQL support
> services. I would avoid adding the extra complexity and additional support
> and maintenance burdens unless you really need it - can work afford to be
> without the system for a day if it crashes? And do you expect it to happen
> infrequently enough that the outages will not be a problem? The advantage of
> having a human available to restore functionality is that they will be able
> to deal with the situations you haven't thought of, whereas an automated
> solution will likely only deal with the situations you have thought of as
> well as making the system more complex, thus creating more things that could
> go wrong.
Right. There was a great 1998 article in the Atlantic Monthly about the
Valujet crash:
http://www.theatlantic.com/issues/98mar/valujet1.htm
Unfortunately that requires a subscription so I pulled a relivant
paragraph:
http://candle.pha.pa.us/main/valujet
The issue is that reducing risk can increase it, as seen with Valujet,
Three Mile Island, and Chernobyl.
--
Bruce Momjian http://candle.pha.pa.us
EnterpriseDB http://www.enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +
---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster
