<br><br><div class="gmail_quote">On Mon, Aug 2, 2010 at 11:54 AM, Oliver Hoffmann <span dir="ltr"><<a href="mailto:oh@dom.de">oh@dom.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
Hi all,<br>
<br>
I have issues with a xen/drbd setup. The problem here is, that the NIC<br>
between the both primary/primary is down for a second once in a while.<br>
<br>
Jul 29 00:07:07 xen1 kernel: bnx2: eth1 NIC Copper Link is Down<br>
Jul 29 00:07:08 xen1 kernel: bnx2: eth2 NIC Copper Link is Down<br>
Jul 29 00:07:10 xen1 kernel: bnx2: eth1 NIC Copper Link is Up, 1000Mbps full duplex, receive & transmit flow control ON<br>
Jul 29 00:07:10 xen1 kernel: bnx2: eth2 NIC Copper Link is Up,1000Mbpsfull duplex, receive & transmit flow control ON<br>
<br>
Hence the storage repository can't be accessed any more and so on.<br>
<br>
At the moment I can't say whether it is a flaw in the driver or HW or<br>
cable or whatever.<br>
<br>
Does some-one had similar problems (with bnx2)? What is the time limit<br>
for drbd when a link dies? Does it makes sense to increase that limit?<br>
<br></blockquote><div><br>What OS are you using? There was a bug in RHEL/CentOS (and I think other OSes, but can't promise that I'm right about that) that was just patched. The bug resulted in occasional loss on connectivity on network connections using the bnx2 driver. See <a href="https://bugzilla.redhat.com/show_bug.cgi?id=587799">https://bugzilla.redhat.com/show_bug.cgi?id=587799</a> for more information.<br>
<br>HTH,<br><br>Ian<br></div></div>