Device timeout / connection unstable

[lynx]

is this normal?

its happening in random, and almost every time I open a new page or refreshing a page.. the connection is dropping/time out…

then in my pfsensebox tty it says

[stephenw10]

No it's not normal.
Do you have any other errors in the logs?

Steve

[lynx]

its happening a lot now  :(

Last 50 system log entries

Jun 16 10:54:27 kernel: fxp1: link state changed to UP
Jun 16 10:54:25 kernel: fxp1: link state changed to DOWN
Jun 16 10:54:25 kernel: fxp1: device timeout
Jun 16 10:54:25 check_reload_status: Linkup starting fxp1
Jun 16 10:54:19 squid[71647]: Squid Parent: child process 72050 started
Jun 16 10:54:14 php: rc.start_packages: The command '/usr/local/etc/rc.d/squid.sh stop' returned exit code '1', the output was '2014/06/16 10:54:09| WARNING: '192.168.200.0/255.255.255.0' is a subnetwork of '0.0.0.0/0.0.0.0' 2014/06/16 10:54:09| WARNING: because of this '0.0.0.0/0.0.0.0' is ignored to keep splay tree searching predictable 2014/06/16 10:54:09| WARNING: You should probably remove '192.168.200.0/255.255.255.0' from the ACL named 'localnet''
Jun 16 10:54:13 squid[65119]: Squid Parent: child process 6206 exited with status 0
Jun 16 10:54:09 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:54:09 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:54:08 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:54:07 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:54:07 php: rc.start_packages: Not calling package sync code for dependency squid of squid because some include files are missing.
Jun 16 10:54:07 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:54:06 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:54:05 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:54:05 php: rc.start_packages: Restarting/Starting all packages.
Jun 16 10:54:02 check_reload_status: Reloading filter
Jun 16 10:54:02 check_reload_status: Starting packages
Jun 16 10:54:02 php: rc.newwanip: pfSense package system has detected an ip change 0.0.0.0 -> 192.168.200.1 … Restarting packages.
Jun 16 10:54:00 php: rc.newwanip: Creating rrd update script
Jun 16 10:54:00 php: rc.newwanip: Resyncing OpenVPN instances for interface LAN.
Jun 16 10:53:58 squid[65119]: Squid Parent: child process 6206 started
Jun 16 10:53:55 php: rc.newwanip: rc.newwanip: on (IP address: 192.168.200.1) (interface: LAN[lan]) (real interface: fxp1).
Jun 16 10:53:55 php: rc.newwanip: rc.newwanip: Informational is starting fxp1.
Jun 16 10:53:55 kernel: pid 65694 (squid), uid 62: exited on signal 6
Jun 16 10:53:55 squid[65119]: Squid Parent: child process 65694 exited due to signal 6
Jun 16 10:53:55 squid[65694]: The url_rewriter helpers are crashing too rapidly, need help!
Jun 16 10:53:54 check_reload_status: Syncing firewall
Jun 16 10:53:54 check_reload_status: Reloading filter
Jun 16 10:53:54 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:53:52 check_reload_status: rc.newwanip starting fxp1
Jun 16 10:53:47 php: rc.linkup: Hotplug event detected for LAN(lan) but ignoring since interface is configured with static IP (192.168.200.1 )
Jun 16 10:53:45 php: rc.linkup: Hotplug event detected for LAN(lan) but ignoring since interface is configured with static IP (192.168.200.1 )
Jun 16 10:53:44 kernel: fxp1: link state changed to UP
Jun 16 10:53:43 check_reload_status: Linkup starting fxp1
Jun 16 10:53:41 check_reload_status: Reloading filter
Jun 16 10:53:41 kernel: fxp1: link state changed to DOWN
Jun 16 10:53:41 kernel: fxp1: device timeout
Jun 16 10:53:41 check_reload_status: Linkup starting fxp1
Jun 16 10:53:41 Squid_Alarm[73112]: Squid has resumed. Reconfiguring filter.
Jun 16 10:53:40 check_reload_status: Reloading filter
Jun 16 10:53:40 Squid_Alarm[71227]: Reconfiguring filter…
Jun 16 10:53:36 Squid_Alarm[68138]: Attempting restart…
Jun 16 10:53:36 Squid_Alarm[68040]: Squid has exited. Reconfiguring filter.
Jun 16 10:53:35 squid[65119]: Squid Parent: child process 65694 started
Jun 16 10:53:32 php: rc.start_packages: The command '/usr/local/etc/rc.d/squid.sh stop' returned exit code '1', the output was '2014/06/16 10:53:27| WARNING: '192.168.200.0/255.255.255.0' is a subnetwork of '0.0.0.0/0.0.0.0' 2014/06/16 10:53:27| WARNING: because of this '0.0.0.0/0.0.0.0' is ignored to keep splay tree searching predictable 2014/06/16 10:53:27| WARNING: You should probably remove '192.168.200.0/255.255.255.0' from the ACL named 'localnet''
Jun 16 10:53:31 squid[66945]: Squid Parent: child process 92491 exited with status 0
Jun 16 10:53:26 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:53:26 php: rc.start_packages: Reloading Squid for configuration sync
Jun 16 10:53:25 php: rc.start_packages: Reloading Squid for configuration sync

[stephenw10]

Hmm, 'device timeout' is a bit generic to be much help really. Could be a hardware incompatibility, could be a buffer problem. I would usually expect to see some other associated error. You could try booting in verbose mode.
Looking at the fxp(4) man page there is a sysctl that counts 'resource not ready' events. Is that counting up? There are other stats there also that might give a clue:

[2.1.3-RELEASE][root@pfsense.fire.box]/root(2): sysctl dev.fxp.1
dev.fxp.1.%desc: Intel 82559ER Embedded 10/100 Ethernet
dev.fxp.1.%driver: fxp
dev.fxp.1.%location: slot=10 function=0
dev.fxp.1.%pnpinfo: vendor=0x8086 device=0x1209 subvendor=0x8086 subdevice=0x0070 class=0x020000
dev.fxp.1.%parent: pci4
dev.fxp.1.int_delay: 1000
dev.fxp.1.bundle_max: 6
dev.fxp.1.rnr: 0
dev.fxp.1.stats.rx.good_frames: 779916
dev.fxp.1.stats.rx.crc_errors: 0
dev.fxp.1.stats.rx.alignment_errors: 0
dev.fxp.1.stats.rx.rnr_errors: 0
dev.fxp.1.stats.rx.overrun_errors: 0
dev.fxp.1.stats.rx.cdt_errors: 0
dev.fxp.1.stats.rx.shortframes: 0
dev.fxp.1.stats.rx.pause: 0
dev.fxp.1.stats.rx.controls: 0
dev.fxp.1.stats.rx.tco: 0
dev.fxp.1.stats.tx.good_frames: 570078
dev.fxp.1.stats.tx.maxcols: 0
dev.fxp.1.stats.tx.latecols: 0
dev.fxp.1.stats.tx.underruns: 0
dev.fxp.1.stats.tx.lostcrs: 4
dev.fxp.1.stats.tx.deffered: 0
dev.fxp.1.stats.tx.single_collisions: 0
dev.fxp.1.stats.tx.multiple_collisions: 0
dev.fxp.1.stats.tx.total_collisions: 0
dev.fxp.1.stats.tx.pause: 0
dev.fxp.1.stats.tx.tco: 0

What hardware is this NIC in? Is it on-board or PCI? Previous similar errors have been down to poorly seated cards.

Steve

Edit: Typo

[lynx]

my NIC is Intel PRO 100+  2pcs both PCI… i dont use the onboard NIC...

Thanks

[cmb]

You skipped the important part - what about the fxp sysctls Steve asked about?

Usually such timeouts are a bad NIC, or a poorly-seated NIC, or on occasion with some systems where the NIC is sharing an IRQ with something else and that something else somehow messes up the NIC.