Intermittant: ADDRCONF(NETDEV_UP): eth0: link is not ready

rickactivemd · 01-19-2012, 05:21 PM

Hello, I have a colocated server with 1and1. The contract is only a couple weeks old and we are already experiencing down time due to some kind of network failure.

My OS and distro is:

Linux 2.6.32-71.29.1.el6.x86_64 #1 SMP x86_64 x86_64 x86_64 GNU/Linux

Every few hours now the server just loses its network connection and is only available via the Serial Console.

The error that I get is:
ADDRCONF(NETDEV_UP): eth0: link is not ready

Here the last past of the dmesg command:

Code:

e1000e 0000:02:00.0: irq 34 for MSI/MSI-X
0000:02:00.0: eth0: (PCI Express:2.5GB/s:Width x1) 00:25:90:57:40:36
0000:02:00.0: eth0: Intel(R) PRO/1000 Network Connection
0000:02:00.0: eth0: MAC: 3, PHY: 8, PBA No: ffffff-0ff
iTCO_vendor_support: vendor-support=0
iTCO_wdt: Intel TCO WatchDog Timer Driver v1.05
iTCO_wdt: Found a Cougar Point TCO device (Version=2, TCOBASE=0x0460)
iTCO_wdt: initialized. heartbeat=30 sec (nowayout=0)
sd 0:0:0:0: Attached scsi generic sg0 type 0
sd 1:0:0:0: Attached scsi generic sg1 type 0
  alloc irq_desc for 18 on node -1
  alloc kstat_irqs on node -1
alloc irq_2_iommu on node -1
i801_smbus 0000:00:1f.3: PCI INT C -> GSI 18 (level, low) -> IRQ 18
ACPI: I/O resource 0000:00:1f.3 [0x1180-0x119f] conflicts with ACPI region SMIO [0x1180-0x119f]
ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver
EXT3 FS on md1, internal journal
EXT4-fs (dm-0): mounted filesystem with ordered data mode
EXT4-fs (dm-1): mounted filesystem with ordered data mode
EXT4-fs (dm-2): mounted filesystem with ordered data mode
Adding 1959920k swap on /dev/sda2.  Priority:-1 extents:1 across:1959920k
Adding 1959920k swap on /dev/sdb2.  Priority:-2 extents:1 across:1959920k
NET: Registered protocol family 10
lo: Disabled Privacy Extensions
ip6_tables: (C) 2000-2006 Netfilter Core Team
nf_conntrack version 0.5.0 (16384 buckets, 65536 max)
ADDRCONF(NETDEV_UP): eth0: link is not ready
e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow Control: None
0000:02:00.0: eth0: 10/100 speed: disabling TSO
ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
eth0: no IPv6 routers present
warning: `in.proftpd' uses 32-bit capabilities (legacy support in use)

A reboot fixes it, but it only stays on for a short while and then goes down again.

What could be the problem and what can I do to troubleshoot?

Thanks.
Rick

castorw · 01-19-2012, 11:22 PM

1, Check cables - check for shortcircuits or interference - barely possible cause of your problem but possible. And question - reboot helps, but does replugging the ethernet cable help?
2, Maybe a network interface card malfunction. Try other NIC.
3, Driver issue, check whether there are no driver updates for your NIC.

And what is that machine running on ? I mean hardware.

rickactivemd · 01-20-2012, 07:22 AM

The server is a:

Intel(R) Xeon(R) CPU E31220 @ 3.10GHz
Quad Core
12GB RAM
1TB HDD

Since the server is new, and with a fresh build of OS, I don't think its a driver problem.

The server techs at 1and1 insist that its probably a network configuration problem, always trying to take as much responsibility off of them as possible. I often get stonewalled with "you should hire a server administrator because we don't support that", usually with a smug and arrogant tone. I try not to get too upset because i've been f&**ing with computers from age 9 through present (35).

The only thing they said they would do is test the memory while the server was disconnected and that it would take hours.

They also insist that it is connected to the switch, but I can't ping out from it due to the network failure.

It is also completely intermittant, as the server has been fine since yesterday when I originally posted.

If you have any more ideas, please let me know.

Thanks.
Rick

madeglobal · 04-14-2012, 06:08 PM

Hi,

I have EXACTLY the same issue with the same specification hardware from 1and1 ... it's almost spooky that I could have posted the same problems such as intermittant loss of the network, same re-boot messages.

Did you ever resolve this issue? What were the results?

1and1 also recommended the same "check memory" test, which would take the server offline for 12+ hours! I suggested that they move my hard disks into another server instead... but no luck yet.

Please do let me know if you found out anything more. Thank you in advance!

rickactivemd · 04-15-2012, 11:58 AM

After I did the research myself and found what was wrong, asked 1and1 one for a resolution. See original inquiry and the response below:

Code:

>
> We just started a new server contract with you a few months ago (xxxxxxx).
>  We noticed immediately that the server was getting disconnected from the
> network.  It has also happened with the other contract on our account.
>  Since then, we have determined that you are releasing a network driver
> that has reported problems with the current version of CentOS that you
> loaded on the server.
>
> Found a similar possible hardware conflict or configuration issue here:
> http://sourceforge.net/tracker/index...02&atid=447449
>
>
> There are many reported issues with e1000e Ethernet driver and CentOS.
>
> e1000e driver version  is 1.2.7-k2, which was the one released with Centos
> 5.6 originally as listed here in the release notes:
> http://docs.redhat.com/docs/en-US/Re...s/drivers.html
>
>
> the driver that should have been released with our version of CentOS (6.1)
> should have been 1.2.20-k2 as listed in this bug report:
> http://bugs.centos.org/view.php?id=5034
>
> the latest version of the driver was released a few days ago is 1.10.6,
> which is about 8 releases ahead of 1.2.20 as listed here:
> http://sourceforge.net/projects/e100...000e%20stable/
>
> Several forums say using the alternative kmod drivers is suggested,
> downloadable from the elrepo repository, here:
> http://elrepo.reloumirrors.net/elrepo/el6/x86_64/RPMS/
>
> We feel you should have one of your server admins update this device driver
> using the most current version or the alternate version on elrepo.
>
> Please let us know when you have done this.
>
> Thanks.
> Rick

And their response:

Code:

Thank you for contacting us.

To solve your problem, we recommend installing a different network driver.  See below on how to install this:

alternative (and newer) driver package could be installed from elrepo.org
rpm --import http://elrepo.org/RPM-GPG-KEY-elrepo.org
rpm -Uvh http://elrepo.org/elrepo-release-6-4...epo.noarch.rpm
yum install kmod-e1000e

This should solve any problems you are having.


If you have any further questions please do not hesitate to contact us.

Hope this helps somebody.
Rick

madeglobal · 04-15-2012, 12:21 PM

Hi Rick,

I think that maybe you are my hero. Next time I see you in a bar, I'll buy you a beer.

Today I backed up all my sites and databases to a remote server, then got 1and1 to swap out ALL the hardware on my server except the disks (long story!) and then re-imaged the server. Finally, I have put back all the sites again. Server has been offline for 18 hours.

However, I strongly suspect that your three lines of code would have fixed this completely. 1and1 are sooooo poor!

(p.s. I made sure that I checked first that my network hardware was the same as yours, and suprise suprise, it was!).

I really hope that 1and1 customers with networks which intermittently drop out find your solution. Thank you so much!

kiransharma755@hotmail.c · 04-11-2016, 11:40 PM

I am having similar issue in my vm created using virtual box in windows 10 host. I did not have such issue until about a month ago, but now any new VM I create, it has this issue.

First of all, this message is displayed in the log in screen. once I get in and start interface, I got the same.

I used standard Intel® Centrino® Advanced-N 6200 AGN driver. Even researching a lot, I am finding any resolution.

FYI, I am not a Linux admin but want to move into Linux.

I would appreciate you help