[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [users] Re: Unkillable process



on Wed, May 23, 2001 at 05:38:55PM -0700, Eric G. Miller (egm2@jps.net) wrote:
> On Wed, May 23, 2001 at 10:23:13AM -0400, MaD dUCK wrote:
> > also sprach Karsten M. Self (on Tue, 22 May 2001 11:29:18PM -0700):
> > > No.  Your memory's going to be released.  But your files might be
> > > scrambled.  I would *not* 'kill -9' my mysqld server.
> > 
> > one of the reasons why i wouldn't run mysql for any reason in the
> > world! unless you don't need a true database backend.
> 
> There's nothing special about mysql in this respect.  Any process that
> has some uncompleted I/O which is abrubtly forced to exit with SIGKILL
> will leave their files in an inconsistent state.  SIGKILL can not be
> caught or handled by the application.  Don't use it unless you really
> have to or you know or don't care how it might affect any files.  Always
> better to try SIGINT first.

A transactional database should be able to recover from an unexpected
shutdown by rolling back the uncommitted transactions.  MySQL sacrifices
robustness for speed in this regard.  Transactional databases won't save
you from, e.g.:  disk damage or malicious intent.

Still, if you need a fast SQL interface, it's a good system, and its
featureset is a strong argument against gratuitously kill -9'ing
arbitrary processes.

-- 
Karsten M. Self <kmself@ix.netcom.com>    http://kmself.home.netcom.com/
 What part of "Gestalt" don't you understand?       There is no K5 cabal
  http://gestalt-system.sourceforge.net/         http://www.kuro5hin.org
   Disclaimer:          http://www.goldmark.org/jeff/stupid-disclaimers/

Attachment: pgpuvebTQhajx.pgp
Description: PGP signature


Reply to: