Topic: New setup--Can't get SME7.4 on the network

[kevincallan]

Hi.  I've installed SME in server/gateway mode many times where is it the top dog on the network.  This is the first time that I have placed an SME box on the network as a single-NIC node in server only mode and can't even get it to ping.  Can somebody tell me if there is a simple thing I am overlooking?  I've searched the forums and docs and see nothing that solves my problem.  I have set the IP address to 10.1.13.250, mask to 255.255.255.0, and gateway to 10.1.13.254 per our sysadm.

So far, I've:

1.  Replaced the Ethernet cable with another.
2.  Tried a different IP address, also unassigned.
3.  Checked the link and activity lights on the NIC and router.  They light and blink during ping.
4.  Set the SME domain to abc.com but thought that might be wrong (Our domain is abc.com.)  I set it to abc.stl and still have trouble.
5.  Checked ifconfig.  Everything looks normal for eth0.

I don't think a ping problem could be a bug so I didn't report it there.

Any ideas?

>>> A new datapoint:  when I boot, the loader pauses at the point where eth0 is initialized.  For a brief moment the [FAILED] message appears but is quickly replaced by the [  OK  ] message.  Does this help at all?

[idp_qbn]

Just some questions:
When you logon locally at the SME box, can you ping the gateway? Or can you run "console" and "test internet access" successfully?

If you can do either of these, then the SME box is "on" the network OK. If so, that means there is a configuration problem somewhere. 
One possibility is there may already be an active address 10.1.13.250? Try assigning another address and see if that fixes the problem. You can also disconnect the SME box and try pinging 10.1.13.250 to see if anything else responds.

Hope this helps

Cheers
Ian

[kevincallan]

Can't ping the gateway.
Testing the internet from the console fails.
I'm to the point where I think it must be a hardware failure.  Everything seems right on the software side.

Thanks.

[Stefano]

Can't ping the gateway.

Thanks.

can you ping other machines than the gateway?

[kevincallan]

All ping attempts to various machines have failed.  Ping to localhost works.

[idp_qbn]

You say you checked for duplicate IP addresses - and even changed the SME IP.
have you got the Netmask correct? Check with another PC just in case they are using a modified netmask (ie NOT 255.255.255.0). Some networks cut the size of the subnet by using a modified netmask (eg 255.255.255.192 allows for 64 hosts, 62 useable  - x.y.z.0 and x.y.z.63 are reserved)
There's more information here http://tinyurl.com/r6xod

You still may be correct and have a dud NIC, the quick way to test that is to replace the NIC and try again, or try the NIC in another box to see if it works. By the way, is the NIC recognised by SME - there are some that casued problems in the past (can't remember which ones, but you can search the forums)

[kevincallan]

So, riddle me this batman. If it is a hardware problem, then why is it that the packet counters keeps increasing?  The Rx and Tx counters increase upon every call to ifconfig.  I now dismiss my theory that it is the hardware and have to believe that it is something else.

[janet]

kevincallan

Basic connectivity from the Console must be established before chasing other reasons.
You have either setup sme wrongly for your network, or your NIC is unsupported or you have bad connections.
Please provide NIC details here for comment.
You can disable the on board NIC and insert a known supported NIC eg Intel card, and try again, run the Configure this server again from the console.

On a workstation on your network run
ipconfig /all
and show us the details to confirm you are using correct IPs, netmask, gateway etc.

Also plug the cable into another notebook and see that the physical cable and outlet do actually work an dteh workstation can get network access.

Unsupported NIC chipsets may only be partially functional so it can be a hardware issue until proven otherwise. If you cannot ping in our out, then there is a basic problem that needs to be sorted first.

[CharlieBrady]

Run "/usr/sbin/tcpdump -i eth0 -n" and see what traffic is actually on the network (which is causing the Rx counters to increment).

[kevincallan]

@mary:

Here is the dump of ipconfig /all for my Windows box.

Code: [Select]

Windows IP Configuration

        Host Name . . . . . . . . . . . . : CTD401
        Primary Dns Suffix  . . . . . . . : c*******.com
        Node Type . . . . . . . . . . . . : Hybrid
        IP Routing Enabled. . . . . . . . : No
        WINS Proxy Enabled. . . . . . . . : No
        DNS Suffix Search List. . . . . . : c*******.com

Ethernet adapter Local Area Connection:

        Connection-specific DNS Suffix  . :
        Description . . . . . . . . . . . : Broadcom NetXtreme 57xx Gigabit Controller
        Physical Address. . . . . . . . . : 00-18-8B-0F-30-B9
        Dhcp Enabled. . . . . . . . . . . : Yes
        Autoconfiguration Enabled . . . . : Yes
        IP Address. . . . . . . . . . . . : 10.1.13.102
        Subnet Mask . . . . . . . . . . . : 255.255.255.0
        Default Gateway . . . . . . . . . : 10.1.13.254
        DHCP Server . . . . . . . . . . . : 10.1.13.254
        DNS Servers . . . . . . . . . . . : 10.1.13.3
                                            10.1.11.3
        Primary WINS Server . . . . . . . : 10.1.11.3
        Lease Obtained. . . . . . . . . . : Friday, April 09, 2010 4:57:41 AM
        Lease Expires . . . . . . . . . . : Saturday, April 10, 2010 4:57:41 AM

[kevincallan]

@mary:

As far as I can tell, my Dell Dimension 2400 has the Broadcom BCM4401 on-board network interface.  This chip is not listed on redhat.com so maybe this is the problem.

[kevincallan]

@CharlieBrady:

When I run "/usr/sbin/tcpdump -i eth0 -n", I get the following output.  This is a segment but the rest looks very similar.

Code: [Select]

<timestamp> arp who-has 10.1.13.51 tell 10.1.13.102
<timestamp> arp reply 10.1.13.51 is-at 00:0b:dc:b5:36:72
<timestamp> IP 10.1.13.102 > 10.1.13.51: icmp 40: echo request seq 64001
<timestamp> IP 10.1.13.51 > 10.1.13.102: icmp 40: echo reply seq 64001

This shows that it works!  My windows machine is the .102 box and my SME machine is the .51 box.  When I ping from my windows box WHILE tcpdump IS RUNNING, I get:

Code: [Select]

Pinging 10.1.13.51 with 32 bytes of data:

Reply from 10.1.13.51: bytes=32 time<1ms TTL=64
Reply from 10.1.13.51: bytes=32 time<1ms TTL=64
Reply from 10.1.13.51: bytes=32 time<1ms TTL=64
Reply from 10.1.13.51: bytes=32 time<1ms TTL=64

When I terminate tcpdump and ping from my windows box, I get:

Code: [Select]

Pinging 10.1.13.51 with 32 bytes of data:

Request timed out.
Request timed out.
Request timed out.
Request timed out.

Ping statistics for 10.1.13.51:
    Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

In addition, if I access http://10.1.13.51/server-manager while tcpdump is running, I have access.  If I terminate tcpdump, I can no longer access server manager from my browser.

[CharlieBrady]

In addition, if I access http://10.1.13.51/server-manager while tcpdump is running, I have access.  If I terminate tcpdump, I can no longer access server manager from my browser.

That sounds like a problem with the driver or the hardware. The NIC is only working when in promiscuous mode. Replace the NIC.

[Marco Hess]

I am certainly not an expert here, but to me this problem smells a bit like a problem I had with VMware clients accessing the local network. Some networking things worked but other things did not.

In the end it had to do with the offloading options in the ethernet driver.

You may be able to configure some of the driver options to solve your problem.

The problem I had was discussed in http://bugs.contribs.org/show_bug.cgi?id=5730