*HCS* half way to the revolution

myke cuthbert cuthbert at fas.harvard.edu
Thu Apr 18 10:24:43 EDT 2002


Of course the point of having clusters is that you don't really care
if you need to take a machine down for a bit for repair because given
proper advance notice and taking it out of round robin, it doesn't
effect anyone.

zeilberg at hcs.harvard.edu [ Apr 17 ]:
> hcs$ uptime
> 21:05  up 180 days, 6 mins,  28 users,  load average: 0.14, 0.47, 0.46
> 
> compare:
> 
> ice$ ( ssh is01 uptime &&
>        ssh is02 uptime &&
>        ssh is03 uptime &&
>     #  ssh is04 uptime &&  # host down
>        ssh is05 uptime && 
>        ssh is06 uptime && 
>        ssh is07 uptime && 
>        ssh is08 uptime && 
>        ssh is09 uptime && 
>        ssh is10 uptime) | awk -- '{days+=$3}END{print "fas cluster up for measly " days " days"}'
> fas cluster up for measly 69 days*
> 
> Now, imagine a beowulf cluster of hcs boxes?





More information about the hcs-discuss mailing list