Re: [OT] Oddities in network information
On Wed, 29 Feb 2012 12:19:30 -0500, Brad Alexander wrote:
> I have a number of Debian boxes connected into a Foundry FastIron II
> switch. This switch has a bunch of 100BaseT ports and 8 1000BaseT ports.
> I have several servers with GigE NICs plugged into the gig ports on the
> switch, but when I look at mii-tool, it shows:
>
> # mii-tool
> eth0: negotiated flow-control, link ok
>
> However, hosts plugged into the 100BaseT ports show
>
> # mii-tool
> eth0: negotiated 100baseTx-FD flow-control, link ok
Compare the results of "mii-tool" with those from "ethtool", IIRC the
former is a bit outdated.
Anyway, here's my output:
stt008:~# mii-tool -v eth0
SIOCGMIIREG on eth0 failed: Input/output error
SIOCGMIIREG on eth0 failed: Input/output error
eth0: negotiated 1000baseT-FD flow-control, link ok
product info: vendor 00:aa:00, model 57 rev 0
basic mode: autonegotiation enabled
basic status: autonegotiation complete, link ok
capabilities: 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD
advertising: 1000baseT-HD 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control
link partner: 1000baseT-FD 100baseTx-FD 100baseTx-HD 10baseT-FD 10baseT-HD flow-control
> The reason for my concern is that I just changed my backuppc box. The
> old box had 100BaseT interfaces, a P4 processor, and 2G of RAM. The new
> box has Gig interfaces, a dual-core Xeon, and 8G of RAM. I have been
> checking backup speeds, and they are not significantly different. On a
> representative host (I have been spot-checking), a full backup took
> about 10 minutes. On the new hardware, backups are taking about 8
> minutes. I'm not a network engineer, but using a transport that is a
> full order of magnitude faster *should* give more than a 20% decrease in
> backup times...Am I wrong here? I guess the Gig-E mii-tool saying
> "negotiated" rather than saying 1000BaseT or something just adds to my
> concerns.
Gigabit is not as fast as we think (with 125 MB/s of max. theorical speed
and can be a real pain when transferring a bunch of gibibytes of data). Also,
the physical transport layer is just another variable of the chain... well,
ensure your card settings are properly set and detected ("dmesg|grep -i eth"
will help).
Greetings,
--
Camaleón
Reply to: