Network card is acting up.

[shredluc]

Hi guys, i have this little problem, and i'm not sure where to look for the solution. Here is the problem: I have a server with onboard Gigabit net cards, and if there is no traffic for a while, ( 5 -10 minutes - i don't have an exact number) then it seems that they go into this "sleep" mode. When anybody tries connecting, be it ssh, or web, it takes about 30 seconds, before the network cards start responding again. I don't think its a hardware issue because the lights blink when recieving packets. It has to be higher up the ladder.

Here is some info:

# uname -a
Linux epsilon 2.6.11-hardened-r15 #3 SMP Fri Sep 9 10:23:21 CDT 2005 x86_64 Intel(R) Xeon(TM) CPU 3.00GHz GenuineIntel GNU/Linux

# lsmod
Module Size Used by
ohci_hcd 22664 0
floppy 63824 0
pcspkr 4824 0
usbhid 28288 0
e1000 89268 0
ehci_hcd 32776 0
uhci_hcd 33184 0
usbcore 126096 5 ohci_hcd,usbhid,ehci_hcd,uhci_hcd

# lspci
0000:00:00.0 Host bridge: Intel Corporation E7520 Memory Controller Hub (rev 09)
0000:00:02.0 PCI bridge: Intel Corporation E7525/E7520/E7320 PCI Express Port A (rev 09)
0000:00:04.0 PCI bridge: Intel Corporation E7525/E7520 PCI Express Port B (rev 09)
0000:00:05.0 PCI bridge: Intel Corporation E7520 PCI Express Port B1 (rev 09)
0000:00:06.0 PCI bridge: Intel Corporation E7520 PCI Express Port C (rev 09)
0000:00:1d.0 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #1 (rev 02)
0000:00:1d.1 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #2 (rev 02)
0000:00:1d.2 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #3 (rev 02)
0000:00:1d.7 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB2 EHCI Controller (rev 02)
0000:00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev c2)
0000:00:1f.0 ISA bridge: Intel Corporation 82801EB/ER (ICH5/ICH5R) LPC Interface Bridge (rev 02)
0000:00:1f.1 IDE interface: Intel Corporation 82801EB/ER (ICH5/ICH5R) IDE Controller (rev 02)
0000:01:00.0 PCI bridge: Intel Corporation 80332 [Dobson] I/O processor (A-Segment Bridge) (rev 06)
0000:01:00.2 PCI bridge: Intel Corporation 80332 [Dobson] I/O processor (B-Segment Bridge) (rev 06)
0000:02:0e.0 RAID bus controller: Dell PowerEdge Expandable RAID controller 4 (rev 06)
0000:05:00.0 PCI bridge: Intel Corporation 6700PXH PCI Express-to-PCI Bridge A (rev 09)
0000:05:00.2 PCI bridge: Intel Corporation 6700PXH PCI Express-to-PCI Bridge B (rev 09)
0000:06:07.0 Ethernet controller: Intel Corporation 82541GI/PI Gigabit Ethernet Controller (rev 05)
0000:07:08.0 Ethernet controller: Intel Corporation 82541GI/PI Gigabit Ethernet Controller (rev 05)
0000:09:0d.0 VGA compatible controller: ATI Technologies Inc Radeon RV100 QY [Radeon 7000/VE]

I'll post more info if needed. Thanks for any thoughts guys.

[commonloon]

A couple of things (you may have already, but...):

1) Does it to the same thing with ping? The reason I ask is that ssh in particular can "Stall" if it needs to regenerate a key or if it is trying to do a reverse lookup and the reverse is failing.

2) Have you checked the BIOS as far as attempting to turn off any sleep related stuff?

3) It looks like you have the e1000 module. Is this just included from the kernel distro or have you emerged e1000 as well to get any updates since? Are there any e1000 messages in dmesg?

[wjholden]

Chances are you can go into your kernel configuration and remove any ACPI and APM stuff you find to solve the problem. I used to have identical problems with my old nForce2 chipset. There's no sense in allowing a power saving utility to turn off the network card(s) of a server.

By the way, nice rig!

[shredluc]

commonloon:

1. It does the same thing with ping. It does not matter if i am pinging from the console out or a workstation in. Same thing with web service. If i try to view a site, it take aprox. 30 seconds to load up the page. After that it responds almost instantly.

2. Checked the bios and all the settings are correct.

3. The e1000 module is from the kernel, and i have not emerged e1000. This is the only output in dmesg that mentions the module:

e1000: eth0: e1000_probe: Intel(R) PRO/1000 Network Connection
e1000: eth1: e1000_probe: Intel(R) PRO/1000 Network Connection
e1000: eth1: e1000_watchdog: NIC Link is Up 100 Mbps Full Duplex <-shame i have to work with a 10/100 switch.


destuxor:
I don't remember any acpi or apm settings that act on the network cards. It's been a couple of weeks though since i compiled the kernel and i don't remember. I'll check that out as soon as i get in front of the thing again.

By the way: Yeah is sure is. Convinced the owner of the company to drop all contractors and bring everything in-house. It's a dualie P4 Xeon 3.0Ghz with HT enabled. It's going to be our production web server. I think they overkilled it though. It's sure nice to have the kernel compile from scratch before i walk 150 feet to get a can of pop and back. When it started going, i almost pooped myself.