motion now unstable since bridge use [Solved]

[pepedog]

Motion was running stabily, weeks at a time without a hickup,
now the pogopig is running as a bridge, it is unstable, connecting to the port returns one frame and it falls over
.
it is started as a service and systemctl -l status motion.service returns
$this->bbcode_second_pass_code('', '>
systemctl -l status motion.service
● motion.service - motion camera
Loaded: loaded (/etc/systemd/system/motion.service; enabled)
Active: failed (Result: signal) since Fri 2014-05-23 10:22:31 BST; 22min ago
Process: 340 ExecStart=/usr/bin/motion (code=exited, status=0/SUCCESS)
Process: 334 ExecStartPre=/usr/bin/fswebcam -r 320x240 -d /dev/video0 -v /dev/null (code=exited, status=0/SUCCESS)
Main PID: 342 (code=killed, signal=SEGV)

May 23 10:20:35 Pogopig motion[342]: [1] 2 length=614400
May 23 10:20:35 Pogopig motion[342]: [1] 3 length=614400
May 23 10:20:35 Pogopig motion[342]: [1] Resizing pre_capture buffer to 1 items
May 23 10:21:00 Pogopig motion[342]: [1] Started stream webcam server in port 7661
May 23 10:21:30 Pogopig motion[342]: [0] Thread 1 - Watchdog timeout, trying to do a graceful restart
May 23 10:22:30 Pogopig motion[342]: [0] Thread 1 - Watchdog timeout, did NOT restart graceful,killing it!
May 23 10:22:30 Pogopig motion[342]: [0] Calling vid_close() from motion_cleanup
May 23 10:22:30 Pogopig motion[342]: [0] Closing video device /dev/video0
May 23 10:22:31 Pogopig systemd[1]: motion.service: main process exited, code=killed, status=11/SEGV
May 23 10:22:31 Pogopig systemd[1]: Unit motion.service entered failed state.
May 23 10:22:31 Pogopig systemd-coredump[358]: Process 342 (motion) dumped core.
')

rebooted a few times and tried it from the nearest computer to the pogopig,
Again you get one frame, then after a while the reported connection gets reset by the server.

again running :-
$this->bbcode_second_pass_code('', '
systemctl -l status motion.service
● motion.service - motion camera
Loaded: loaded (/etc/systemd/system/motion.service; enabled)
Active: active (running) since Fri 2014-05-23 11:05:19 BST; 3min 19s ago
Process: 398 ExecStart=/usr/bin/motion (code=exited, status=0/SUCCESS)
Process: 392 ExecStartPre=/usr/bin/fswebcam -r 320x240 -d /dev/video0 -v /dev/null (code=exited, status=0/SUCCESS)
Main PID: 400 (motion)
CGroup: /system.slice/motion.service
└─400 /usr/bin/motion

May 23 11:07:17 Pogopig motion[400]: [1] found control 0x00980913, "Gain", range 0,255
May 23 11:07:17 Pogopig motion[400]: [1] "Gain", default 0, current 16
May 23 11:07:17 Pogopig motion[400]: [1] mmap information:
May 23 11:07:17 Pogopig motion[400]: [1] frames=4
May 23 11:07:17 Pogopig motion[400]: [1] 0 length=614400
May 23 11:07:17 Pogopig motion[400]: [1] 1 length=614400
May 23 11:07:17 Pogopig motion[400]: [1] 2 length=614400
May 23 11:07:17 Pogopig motion[400]: [1] 3 length=614400
May 23 11:07:17 Pogopig motion[400]: [1] Resizing pre_capture buffer to 1 items
May 23 11:07:39 Pogopig motion[400]: [1] Started stream webcam server in port 7661
')
I know it s just one frame as its got the real time clock in the bottom corner

any ideas where to look to see why it looks like the webstream server is stopping and why motion keeps crashing ?
TIA

[pepedog]

So when you restore it to not-as-a-bridge it is fine? I can't see how the two things connected, unless it binds to an ip address or Ethernet interface

[g8jvm]

I haven't tied going back pepedog,
But as the posty delivered a new USB/SATA caddy today
I've done a fresh install onto the hard drive. 230GB a spare new drive I had.

Again I found exactly the same behavior with addresses, it did not want the onboard NIC MAC address for br0 , but the USBNIC.
I've got all three partitions mounted, and apart from samba, its a clean install with just the required packages for bridge and samba.

my webcam port is 7661, and what ever browser I use to connect it gets the connection refused,
systemctl status motion.service shows
$this->bbcode_second_pass_code('', '
[root@pogopig ~]# systemctl status motion.service
● motion.service - motion camera
Loaded: loaded (/etc/systemd/system/motion.service; enabled)
Active: failed (Result: signal) since Fri 2014-05-23 16:04:57 BST; 13s ago
Process: 347 ExecStart=/usr/bin/motion (code=exited, status=0/SUCCESS)
Process: 338 ExecStartPre=/usr/bin/fswebcam -r 320x240 -d /dev/video0 -v /dev/null (code=exited, status=0/SUCCESS)
Main PID: 350 (code=killed, signal=SEGV)

May 23 16:03:09 pogopig motion[350]: [1] 2 length=614400
May 23 16:03:09 pogopig motion[350]: [1] 3 length=614400
May 23 16:03:09 pogopig motion[350]: [1] Resizing pre_capture buffer to 1 items
May 23 16:03:28 pogopig motion[350]: [1] Started stream webcam server in port 7661
May 23 16:03:57 pogopig motion[350]: [0] Thread 1 - Watchdog timeout, trying to do a graceful restart
May 23 16:04:57 pogopig motion[350]: [0] Thread 1 - Watchdog timeout, did NOT restart graceful,killing it!
May 23 16:04:57 pogopig motion[350]: [0] Calling vid_close() from motion_cleanup
May 23 16:04:57 pogopig motion[350]: [0] Closing video device /dev/video0
May 23 16:04:57 pogopig systemd[1]: motion.service: main process exited, code=killed, status=11/SEGV
May 23 16:04:57 pogopig systemd[1]: Unit motion.service entered failed state.
')

I've run top and watched the usage while connecting to the webcam server, yet no big usage shown

The motion istall is exactly as per
blog.qnology.com/2013/03/tutorial-pogoplug-e02-with-arch-linux.html

which I was using before.
It was all tested and running on the previuos install on a 8GB mememory stick.

When bridge configuration is used is any form of firewall or port blocking configured. ?

while typing I've rebooted and I'll look now with a browser, Firefox, for the webcam

$this->bbcode_second_pass_quote('', 'F')irefox can't establish a connection to the server at 192.168.51.5:7661.

[pepedog]

Is port 7661 a http port? Or pipe?
If http does it bind to an interface?

[g8jvm]

Hi
I'm guessing its a http port, as I've always used http to access it,

not sure if its bond to a port, how do I check that ?

[pepedog]

Not bond to a port, bind to an interface.
Look in conf file for eth0

[g8jvm]

/etc/netctl/eth0
$this->bbcode_second_pass_code('', '
[root@pogopig netctl]# cat eth0
Description='A basic dhcp ethernet connection'
Interface=eth0
Connection=ethernet
IP=dhcp
## for DHCPv6
#IP6=dhcp
## for IPv6 autoconfiguration
#IP6=stateless

ExecUpPost='/usr/bin/ntpd -gq || true'
DhcpcdOptions='-A'
ExecUpPost='/usr/bin/ntpdate -u pool.ntp.org'
')

thats the only thing with eth0 in the file name except for
[root@pogopig interfaces]# locate eth0
/etc/netctl/eth0
/var/lib/dhcpcd/dhcpcd-eth0.lease


gdb says the reason for the core dump is a segmentation fault.

[pepedog]

Not that
/etc/motion.conf, or whatever it's called

[g8jvm]

Its something to do with the USB ports, If I reduce the power loading on the ports , motion no longer segfaults.
there was a USB repeater on the webcam line, if I remove the repeater, the port loading drops , still has the same amount of devices.
less the repeater.
That was in use before bridging with no problems, the only difference was the USBNIC was plugged in so powered up, but the network on that port was down.
Damn wierd .
well both working now

thanks Pepedog