jhcarter Posted October 30, 2015 Posted October 30, 2015 I am running Armbian_4.5_Cubox-i_Ubuntu_trusty_3.14.54 on a CuBox i4 Pro. The application is a security camera server and the current test environment consists of the server and two IP cameras set to 320x240 and running at 5 fps. The server is receiving these two streams over the Ethernet interface constantly but the data rate is pretty low. I am seeing the Ethernet interface fail after 24 to 48 hours with the syslog entries shown below. After the failure I can restart the link with ifdown/ifup and it works fine. Has anyone else seen a similar issue? Oct 29 11:56:09 localhost kernel: [ 0.505433] libphy: fec_enet_mii_bus: probed Oct 29 11:56:09 localhost kernel: [ 0.505892] fec 2188000.ethernet eth0: registered PHC device 0 Oct 29 11:56:09 localhost kernel: [ 3.832211] fec 2188000.ethernet eth0: Freescale FEC PHY driver [Atheros 8035 ethernet] (mii_bus:phy_addr=2188000.ethernet:04, irq=-1) Oct 29 11:56:11 localhost kernel: [ 6.837655] libphy: 2188000.ethernet:04 - Link is Up - 100/Full Oct 30 03:17:01 localhost kernel: [55253.929843] fec 2188000.ethernet eth0: MDIO read timeout Oct 30 03:17:03 localhost kernel: [55255.940256] libphy: 2188000.ethernet:04 - Link is Down
Igor Posted November 1, 2015 Posted November 1, 2015 No, never saw this. I use one Cubox on daily basis but with a little different set-up (Openelec) but it works w/o problem. I think the Ethernet driver is the same. I am not sure if there are some changes in this area but just in case try with some older kernel 3.14.14 http://apt.armbian.com/pool/main/l/linux-upstream/linux-image-cubox_4.2_armhf.deb
jhcarter Posted November 1, 2015 Author Posted November 1, 2015 I will try the 3.14.14 kernel then. It will take few days to do a proper test. I have found someone else reporting that this issue was corrected by installing a newer kernel. Unfortunately he does not state which kernel versions were involved. http://superuser.com/questions/913075/wired-connection-being-dropped Thanks Igor.
jhcarter Posted November 20, 2015 Author Posted November 20, 2015 This happened two times over a few days just prior to my initial post. In an effort to be sure I have a repeatable failure test case I have let it run since that time with few reboots and with the original 3.14.54 kernel and it has not failed again. I must assume at this point that there was some external influence at work with the initial two failures.
Recommended Posts