Hi,
I have problem with a Centos 5.3 computer. The networking is very slow. The networking card is a RealTek 1GigE.
I plugged an Ubuntu laptop on the same cable and port with an Intel GigE, and it's about 8 to 10 times faster. So the problem is likely to be with the CentOS config.
Any suggestion on how to track down the problem with speed.
I have checked the packets with wireshark, no problem.
Here's my ethtool eth0 output:
Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full
Here's ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:06:7B:00:14:16 inet addr:192.168.2.103 Bcast:192.168.2.255 Mask:255.255.255.0 inet6 addr: fe80::206:7bff:fe00:1416/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:3139398 errors:0 dropped:0 overruns:0 frame:0 TX packets:2890605 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:4048643591 (3.7 GiB) TX bytes:257250846 (245.3 MiB)
Any suggestion?
lspci -v output whould be helpful
centos@911networks.com wrote:
Hi,
I have problem with a Centos 5.3 computer. The networking is very slow. The networking card is a RealTek 1GigE.
I plugged an Ubuntu laptop on the same cable and port with an Intel GigE, and it's about 8 to 10 times faster. So the problem is likely to be with the CentOS config.
Any suggestion on how to track down the problem with speed.
I have checked the packets with wireshark, no problem.
Here's my ethtool eth0 output:
Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full
Here's ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:06:7B:00:14:16 inet addr:192.168.2.103 Bcast:192.168.2.255 Mask:255.255.255.0 inet6 addr: fe80::206:7bff:fe00:1416/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:3139398 errors:0 dropped:0 overruns:0 frame:0 TX packets:2890605 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:4048643591 (3.7 GiB) TX bytes:257250846 (245.3 MiB)
Any suggestion?
ethtool eth0 and lsmod
would be helpful as well.
centos@911networks.com wrote:
Hi,
I have problem with a Centos 5.3 computer. The networking is very slow. The networking card is a RealTek 1GigE.
I plugged an Ubuntu laptop on the same cable and port with an Intel GigE, and it's about 8 to 10 times faster. So the problem is likely to be with the CentOS config.
Any suggestion on how to track down the problem with speed.
I have checked the packets with wireshark, no problem.
Here's my ethtool eth0 output:
Advertised auto-negotiation: Yes Speed: 1000Mb/s Duplex: Full
Here's ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:06:7B:00:14:16 inet addr:192.168.2.103 Bcast:192.168.2.255 Mask:255.255.255.0 inet6 addr: fe80::206:7bff:fe00:1416/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:3139398 errors:0 dropped:0 overruns:0 frame:0 TX packets:2890605 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:4048643591 (3.7 GiB) TX bytes:257250846 (245.3 MiB)
Any suggestion?
centos@911networks.com wrote:
Hi,
I have problem with a Centos 5.3 computer. The networking is very slow. The networking card is a RealTek 1GigE.
Get a better NIC, Realtek is absolute crap. CentOS is an "enterprise grade" OS, use an "enterprise grade" NIC such as Intel or Broadcom.
nate
Get a better NIC, Realtek is absolute crap. CentOS is an "enterprise grade" OS, use an "enterprise grade" NIC such as Intel or Broadcom.
nate
+1 Realtek NICs are well know for being buggy in Linux specialy when a lot packets are involved.
nate wrote:
centos@911networks.com wrote:
Hi,
I have problem with a Centos 5.3 computer. The networking is very slow. The networking card is a RealTek 1GigE.
Get a better NIC, Realtek is absolute crap. CentOS is an "enterprise grade" OS, use an "enterprise grade" NIC such as Intel or Broadcom.
nate
+1
What chipset are they?
I've had poor results with cards featuring the 8169 chip.
What kind of switch is it plugged into? If it is managed, can you check the switch to see what was negotiated? I've seen autoneg issues a lot on certain brands of switches (Nortel e.g.)
Hi,
On Wed, Sep 16, 2009 at 11:25, centos@911networks.com wrote:
I plugged an Ubuntu laptop on the same cable and port with an Intel GigE, and it's about 8 to 10 times faster. So the problem is likely to be with the CentOS config.
How did you test that? Opening a web page? Downloading a big file?
Any suggestion on how to track down the problem with speed.
First check if it's a DNS issue, check your /etc/resolv.conf to see if both DNS servers there are correct, then check with "host" or "nslookup" querying all of them to see if you get replies all of the time.
If you have an iptables firewall running, you might want to try to temporarily disable it to see if there is a difference.
If it's not DNS or Firewall, you might want to use "iperf" to measure the bandwidth available between that host and another host that works well in your network. You can get "iperf" from the RPMforge repository (http://wiki.centos.org/AdditionalResources/Repositories/RPMForge) or from http://iperf.sourceforge.net/ (binaries for Linux and other platforms should be available). From "iperf" you can really know if there is an actual difference from your CentOS machine to your Ubuntu laptop, with those numbers we might be able to give you better suggestions on how to diagnose your problem.
HTH, Filipe