>-----Original Message----- >From: Jason Keltz [mailto:j...@cse.yorku.ca] >Sent: Friday, July 03, 2015 11:19 AM >To: e1000-de...@lists.sf.net >Subject: [E1000-devel] dropped packets on ixgbe centos 7.1 > >Hi. > >I have 3 Dell R620 systems that have dual Intel 10G X540-AT2 running >CentOS 7.1 and oVirt virtualization software. >One 10G port on each system is dedicated to the virtualization >management network and is connected to a Netgear XS712T 10G switch. >The CentOS servers are running the default RHEL7.1 ixgbe driver: >Intel(R) 10 Gigabit PCI Express Network Driver - version 4.0.1-k-rh7.1 > >I'm trying to solve two problems, and I wonder if they are related: > >First, I notice that there is a significant number of dropped RECEIVED >packets on each system:
The ethtool stats reported by the driver (ethtool -S) do not show any dropped packets. For dropped packets you usually look at the following counters: rx_errors: 0 rx_missed_errors: 0 tx_flow_control_xoff: 0 rx_no_dma_resources: 0 <snip> >Now, if I can solve this problem, there's the second problem, and it may >or may not be related... > >Each of the servers has idrac7 Express - the Dell Remote Access >Controller. Because it's express, and not "enterprise", it shares one >of the system network ports, which happens to be the management >network. I'm having a problem where-by, after some period of time >(could be a week, could be a month), the Dell Remote Access controller >stops responding to ping entirely (100% packet loss). I thought this >was unrelated to the 10G nic. If I hard reboot the idrac (with an ipmi >command under Linux), it reboots, but following the reboot, pinging it >drops 80-90% of all packets. It's the same every time the problem >occurs. Meanwhile, I can still ping the underlying management >interface, and all my packets get through. But here's the really weird >part -- if I reboot the SERVER (and not the idrac - the idrac runs >independent from the server of course), then the drac starts responding >to ping again, with no packet loss!! Dell hasn't been able to replicate >the problem in house, but to date, they're been working with idrac >Enterprise (which has a separate network port), and even if they were >working with the express, there's the issue of bridge versus >non-bridge. My virtualization servers have all got into this state. >The file server, which doesn't use the bridge, hasn't got into this >state... The virtualization servers all show dropped packets... the file >server doesn't... related???? Considering the packets are not dropped by the interface I don't think it's related. Sounds like you have already reported the issue to Dell which is the correct way to go about it. Thanks, Emil ------------------------------------------------------------------------------ Don't Limit Your Business. Reach for the Cloud. GigeNET's Cloud Solutions provide you with the tools and support that you need to offload your IT needs and focus on growing your business. Configured For All Businesses. Start Your Cloud Today. https://www.gigenetcloud.com/ _______________________________________________ E1000-devel mailing list E1000-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/e1000-devel To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
