Re: pgsql: Improve wording. - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: pgsql: Improve wording.
Date
Msg-id 200712220519.lBM5JND26496@momjian.us
Whole thread Raw
List pgsql-hackers
bruce wrote:
> > I think your proposed wording is removed enough from what the
> > complainant was saying that it is not worth to stick it in.  The point
> > here is, to what extent do we want to spoon-feed careless sysadmins?
> 
> OK, I have removed the paratheses paragraph about fork() and added link
> to an article about the OOM:
> 
>     http://lwn.net/Articles/104179/
> 
>     On Linux 2.6 and later, an additional measure is to modify the
>     kernel's behavior so that it will not <quote>overcommit</> memory.
>     Although this setting will not prevent the <ulink
>     url="http://lwn.net/Articles/104179/">OOM killer</> from being invoked
>     altogether, it will lower the chances significantly and will therefore
>     lead to more robust system behavior.  This is done by selecting strict
>     overcommit mode via <command>sysctl</command>:

I just read that article in full and it does a better job of explaining
OOM than we ever could: :-)
   Not everybody agrees that this (OOM) is a fruitful use of developer time.   Andries Brouwer came up with this
analogy:Anaircraft company discovered that it was cheaper to fly its planeswith less fuel on board. The planes would be
lighterand use less fueland money was saved. On rare occasions however the amount of fuel wasinsufficient, and the
planewould crash. This problem was solved by theengineers of the company by the development of a special
OOF(out-of-fuel)mechanism. In emergency cases a passenger was selected andthrown out of the plane. (When necessary, the
procedurewas repeated.) Alarge body of theory was developed and many publications were devoted tothe problem of
properlyselecting the victim to be ejected. Should thevictim be chosen at random? Or should one choose the heaviest
person?Orthe oldest? Should passengers pay in order not to be ejected, so thatthe victim would be the poorest on board?
Andif for example theheaviest person was chosen, should there be a special exception in casethat was the pilot? Should
firstclass passengers be exempted? Now thatthe OOF mechanism existed, it would be activated every now and then,
andejectpassengers even when there was no fuel shortage. The engineers arestill studying precisely how this malfunction
iscaused.
 


--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://postgres.enterprisedb.com
 + If your life is a hard drive, Christ can be your backup. +


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: pgsql: Improve wording.
Next
From: Tom Lane
Date:
Subject: Re: timetz range check issue