From: Holger.VanKoll@SWISSCOM.COM
Date: Fri Jul 12 2002 - 13:09:37 EDT
use ftp without disc-access to measure better (without encrpytion->cpuload , without disk-caching issues)
I use this .netrc
default
login your_username
password your_password
macdef init
bin
put "|dd if=/dev/zero bs=1048576 count=100" /dev/null
quit
post the no -a output from the nodes that behave differently/bad
> -----Original Message-----
> From: Jolet, John [mailto:John.Jolet@MISYSHEALTHCARE.COM]
> Sent: Freitag, 12. Juli 2002 18:47
> To: aix-l@Princeton.EDU
> Subject: transfer problem
>
>
> Okay, here's an odd one. I've got an H50 that is acting funny....for
> instance...i transferred a 26gig file TO it in 3 minutes. To
> transfer that
> file FROM it...well, scp was estimating over an hour. What
> do I look at
> first? it can receive fine, but not transmit?
> To fully qualify the program....file transmitted from machine
> a to b took 3
> minuts. from b to c stopped it at 15, 20% done. From a to c 4
> minutes...from b to d 2 minutes....from c to b, 2
> minutes...pulling from b
> to c (scp from c, but with source on b) same result as
> pushing from b to
> c...pulling from b to a...37 seconds.
>
> Why can this machine not transmit properly to that ONE box?
> but can receive
> from that box just fine?
>
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 22:16:03 EDT