Hot answers tagged

38

Ok so after posting this question last night night I continued to do some research the only real solution I came across seems to have taken care of the problem. Disabling TSO, GSO and GRO using ethtool: ethtool -K eth0 gso off gro off tso off According to a post found here: http://ehc.ac/p/e1000/bugs/378/ From what I understand this will or can cause a ...


16

ethtool -l <interface> will display the status of queues associated with an interface, if that interface's driver supports such a thing. In ethtool-land, multiqueue is indicated by "channels". If you see responses from ethtool like: homeserver-02 ~ # ethtool -l enp4s0 Channel parameters for enp4s0: Cannot get device channel parameters : Operation ...


15

These devices are typically seen in the OS as four separate network devices and they can be used just like the four separate NICs you had before.


13

Yes, they're completely hot-swappable. Think of them the same way as cables (note the link lights). Same for SFP+ direct-attach cables. It wouldn't make sense for them not to be hot-swappable. Heck, most network switches don't even have power switches :)


12

If your NICs have the same MAC addresses, you should stop fiddling around with workarounds and return them to the manufacturer as defective. Get proper replacements and continue on as normal. Incur the downtime once instead of the recurring issues that will pop up from continuing to hack together a "solution."


12

A "network device" (in the context of a computer) typically refers to a piece of hardware - the network interface card (NIC). Once upon a time, back when dinosaurs roamed the internet and lots of people still used modems, NICs really only came with one interface, but nowadays, NICs with multiple interfaces are common. A NIC can have multiple active IP ...


12

You don't want to buy a "fiber card". What you want is a card (or preferably a switch) that has either SFP or SFP+ ports. This will allow you to purchase cheap optics depending on what you need, and will ensure that you're future-proof if you need to upgrade or change optics in the future. As for what type of optics you'll need, that's a conversation you ...


10

Having configured a number of SmartUPSes with AP9606 in my time, I do not remember ever having had the need for an "APC-proprietary" serial cable. The documented pinout looks like a simple null-modem cable swapping the TX and RX pins: +-----------------------------------------------------------------+ | ...


10

check for errors on the wire, look at the "errors" field in the output of ifconfig. If non-zero then there are problems with hardware (cable, NIC card, or hub/switch). An unreliable Ethernet cable will give errors in this field too. replace the Ethernet cable, regardless of step 1. This is quick, cheap and easy, and should be done whenever your link is going ...


10

Long story short: remove the virtual switch from Hyper-V, recreate it, and everything worked again...


9

Disabling Enhanced C1 (C1E) in the BIOS fixed it for me. Not sure if the lower power state of C1E is messing with the driver, or that there's an oops in the driver when the processor is in this state. Anyway, problem solved.


8

For any network device which is not my default gateway, I usually set the default route flag to no: DEFROUTE="no" This seems to work without any issues for my servers with multiple network interfaces. If you then restart your network service or interfaces, you should be able to check the routes to see that this is actually working: /sbin/route -n ...


7

In general, you don't do the bonding on a virtual machine, you do it on the hypervisor. You then present the bonded interface to the virtual machines as a single interface.


7

You more likely want to create a bridge with the 8/9 interfaces and then assign an IP address to that bridge (bridge-utils packet, command 'brctl add'). This way the bridge will act like as a switch and can have an IP address into your subnet.


7

Generally yes they are, certainly I've never come across any that weren't - in particular I know for a fact that the top three on your list are, not used the bottom one so can't say with the same degree of confidence.


7

Well, there's the example of the pre-set maximum ring buffer figures on Broadcom bnx2 devices being modified in the kernel from 1020 to 2040 a few years ago, so it is possible. diff --git a/drivers/net/bnx2.h b/drivers/net/bnx2.h index efdfbc2..62ac83e 100644 --- a/drivers/net/bnx2.h +++ b/drivers/net/bnx2.h @@ -6502,8 +6502,8 @@ struct l2_fhdr { #define ...


7

Under the assumption you are talking about users using Remote Desktop Services. Windows Server 2008 R2 has a feature called Remote Desktop IP Virtualization and comes with the Remote Desktop Server Session Host role. This feature comes with 2 options: Per-Session mode: Remote Desktop IP Virtualization assigns an IP address per user session. Per-program ...


6

Well, it may be proprietary (in the sense of it being a non standard pinout), but you can find pinout diagrams if you know what you are looking for, and if you have a few inexpensive tools, you can also make one. Here's a fairly simple one: This is the 0024C model: +-----------------------------------------------------------------+ | ...


5

The easiest way to do this is to use a custom-built cable with doesn't have the TX pin connected. Cables are trivial to make, you just have to follow the correct color pattern. You'll also have to make sure you put the TX-disabled connector in the computer and not the switch, but that's easy to fix if you get it wrong. :)


5

You should examine/post any errors you've received... Why do you suspect you have a problem? What behaviors have you seen on the system. And you should definitely have multiple NICs available for hypervisor VM traffic to counter this issue. If you suspect the NIC is bad, and it's a standalone device, replace it. If it's on a system board/motherboard, check ...


5

In Red Hat 5 and CentOS 5, you need to install VMware Tools, because the vmware modules aren't include the main line kernel. For example, in SUSE 11 Sp1 with kernel 2.6.32.59-0.7-default: grep -i vmxnet3 /boot/config-2.6.32.59-0.7-default CONFIG_VMXNET3=m as you can see the module is integrated in the kernel as external module, anyway you can do a test ...


5

You can use a single port NIC and a dual-port NIC... or just use a 4-port card. There's no real history of 3-port network add-in cards. I'm sure you can find something available in your price range. This is a commodity component.


5

This is an important question and I'd say a more common scenario than it appears from your searching. As you may know there are three types of teaming provided by MS Server. 1. Active / Standby 2. Static 3. LACP Based on your statement about whether you will have to tear everything down it sounds to me like you are using Static teaming which ...


5

Yes. Yes. Yes. It can be but doesn't by default. When the packet arrives its destination is evaluated by its internal virtual switch to decide which VMs to send it to or if it's for its own kernel.


5

As others have already mentioned, yes, a 4-port network card will be recognized and usable just like 4 separate network interfaces. However, in particular with 4-port cards, you should pay attention to the bus bandwidth the card is capable of. For example, the Intel EXPI9404VT in your second link has a PCIe 1.1 x4 bus interface, meaning that it has a total ...


4

I know, in theory, how a hardware interrupt works, but how does it specifically function when a NIC receives packets of information? What is happening on a hardware level? When the NIC receives information, it checks to see if the conditions are met to trigger a hardware interrupt. This is typically done in firmware on the NIC controller. If, for example, a ...


4

According to this link: http://www.intel.com/support/network/adapter/pro100/sb/CS-032498.htm These are server based nic's. Having said that, some desktop systems use the the igb driver as well. The 82576 is hardly aimed at desktop use. It's stated to be for dual port NICs and virtualization. What NIC model/chipset do you have? Edit: You've stated you ...


4

One of the problems with high performance NIC's is that the modern PC architecture has a bit of trouble keeping up. But, in your case, this isn't so much the problem. Let me explain. The CPU has to do a lot of work processing TCP packets. This affects the throughput. What's limiting things in your case is not the network hardware, but the ability of the ...


4

A "bridge" is a layer 2 phenomenon, which won't help you communicate between two different subnets. What you have is a route (and firewall rules that permit traffic between the two subnets).


4

Disabling only TCP Segmentation Offload (TSO) does the trick for me. ethtool -K eth0 tso off Note: It does not seem to be necessary to also disable Generic Receive Offload (GRO) and Generic Segmentation Offload (GSO), as it is recommended by various sources. As far as I learned, these are implemented purely in software, and should be safe. Don't sacrifice ...


Only top voted, non community-wiki answers of a minimum length are eligible