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

RE: computer freezes

This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.

  • Subject: RE: computer freezes
  • Date: Fri, 6 May 2005 11:10:44 -0500 
 
>RedHat and/or Intel just can't handle a high load average.  A load
>average of
>10 is sufficient to totally hose a RH system, possibily requiring a
>reboot.
>I've seen load averages > 10 on other systems, without user performance
>being
>affected.

Since this was brought up, I am only replying to this point only.

IMO this is definitely Linux behavior and not the Intel platform.  This has
nothing to do with it being an Intel based system.  It's Linux that can't
handle such a high load average.  This is one reason why I think it is
unsuitable if you REALLY want to run for a year without rebooting, if you
work it hard.  The fact that people put up with this as the norm is
mystifying to me.  SPARC/Irix/PowerPC systems don't handle high loads better
necessarily, the OSes that run on them do,

Solaris x86 has no such issues.  I have two development Solaris x86 LDM
boxes (Dell 1400SC) that had some processes that ran amuck (my fault) that
caused load averages to be over 14 for a week and people continued to get
data from them.  They noticed things were a tad slow, but nobody thought
there was an issue until I happened to check on them and saw the runaway
processes.

Linux may be cool, and for some things it may be faster, but if you truly
push it hard it will fail you every time.  Since many people see this as a
religious issue, I will not blast the whole group.