2 messages in net.nether.puck.cisco-nsp[c-nsp] Catalyst 3750G + Broadcom NIC...
FromSent OnAttachments
Tristan GulyasJan 10, 2005 5:29 am 
Colin WhittakerJan 10, 2005 6:24 am 
Actions with this message:
Paste this link in email or IM:
Paste this link in email or IM:
Atom feed for this thread
Paste this URL into your reader:
Subject:[c-nsp] Catalyst 3750G + Broadcom NIC issuesActions...
From:Tristan Gulyas (zar@hotblack.net)
Date:Jan 10, 2005 5:29:10 am
List:net.nether.puck.cisco-nsp

Howdy all,

I've run into a little issue of late with Cisco 3750G-24T and -24TS switches and Broadcom NeXtreme Gigabit controllers supplied onboard the Compaq d530 small form factor systems under Linux.

The issue: machines lose link and the NIC freezes up after a few days of inactivity. Reloading the module (which shuts down the controller and restarts it) seems to help. I even had an issue the other day where the switch would not provide link at all unless the port was set to 100mbit and that didn't last long, either.

Cabling is freshly unpacked Cat5e - I've had no troubles with the same cable drop on my PowerBook G4.

These problems are not present on the 6500 10/100 or Gigabit-capable blades we also connect to, nor the 10Mbit HP hubs we have lying around.

I've tried manual and automatic settings but with no avail. Reloading the switch doesn't help.

I've seen the same results with 12.2(2)SE3, a beta of 12.2(25)SE and 12.1(19)EA1.

Configuration from the switch ports is as follows:

Linux console simply reports:

tg3: eth0: Link is down.

Card in kernel reports:

eth0: Tigon3 [partno(BCM95782A50) rev 3003 PHY(5705)] (PCI:33MHz:32-bit) 10/100/1000BaseT Ethernet 00:0e:7f:2a:34:43

No changes occur at either end that seem to trigger this.

Any help or suggestions would be appreciated.

thanks,

.t

PS. If anyone has used any of the new Fluke network assistant devices with these switches and has had issues obtaining link for more than 2 seconds, let me know...