Film-Tech Forum ARCHIVE: Datasat AP20 malfunctioning ethernet interface...

This is topic Datasat AP20 malfunctioning ethernet interface... in forum Digital Cinema Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=16;t=002878

Posted by Christian Hove (Member # 6743) on 11-29-2016, 06:28 AM:

Hi,
One of our Datasat AP20s exhibits erroneous behavior at least three times during the last two weeks - The ethernet interface stops working. After powercycling the unit (plug-out-wait-plug-in) the interface works again.

Using system->network->netdiag I see:

From this I conclude:
- The physical link is OK
- The ethernet interface has detected incoming malformed frames

Now I try two things:
1. Ping from the network to the AP20
2. Ping from the AP20 to the network (gateway)

1. Pinging the AP20 from the network generates no answer at the pinging host and additional Rx frame errors at the AP20.

2. Pinging the gateway from the AP20 generates an error message AND no increased Tx packet count!

Enabling Net stats logging, I see the same thing browsing through the log - Pinging from the AP20 does not generate increased Tx packet count as it does on our other AP20s.

I have tried the obvious; replacing ethernet cable and switch with no effect. I have also upgraded to FW V1.05.01 with no effect either.

Any ideas on how to proceed on this one?

Posted by Dave Macaulay (Member # 813) on 11-29-2016, 07:16 AM:

I don't know what a "syberr" is but it sounds bad. Definitely though, dropping 1/5 of the rx packets is double plus ungood.
If you've replaced the cabling and the switch, either the port hardware is failing on the AP or something it talks to is sending garbage. Since the port shuts down intermittently it's likely a fault in the AP. It's unusual for ports to shut down because of bad rx data, but consumer grade switches can stop working when hit with some kinds of corrupt traffic.

Posted by Marco Giustini (Member # 4544) on 11-29-2016, 11:45 AM:

IP conflict? Can you just wire your laptop into the AP20 and send a continuous ping and see what happens?

Posted by Carsten Kurz (Member # 5396) on 11-29-2016, 05:08 PM:

Can you backup this and another of your working AP20 and swap both?

Long shot, maybe the power supply? Can you compare the voltage readings between this and another unit?

- Carsten

Posted by Leslie Hartmier (Member # 7053) on 11-29-2016, 07:58 PM:

syberr is a subroutine error handler for Sybase in Perl. It's used to set an ad-hoc error handler callback that gets called before the normal error handler does its job. If this subroutine returns 0 then the error is ignored.

The above is informative, but is it useful? I know, NO.

What is useful is that you should never see anything but zero on that.

The fact that you do might make one concerned. That said, this is what I would do to troubleshoot:

1) add a known good router to the AP20, and ONLY the AP20 and router are connected (well, and a laptop/computer, of course),
2) switch the AP20 to DHCP,
3) let the router give the AP20 an IP address (check the router's DHCP client page to see if the AP20's MAC address is there, and what the router gave the AP20 for an IP address),
4) check the AP20's network page to see what the IP address for the AP20 is to compare to what you see on the DHCP client page on the router,
5) then try to ping it using a machine on the same network.

The reason I recommend this is that a direct connection to the router takes away any chance that there is something in the way, that if the AP20 receives an IP address, then you know that the DHCPDiscover message is being sent to the local subnet created by the router (tells you you have connectivity to an external device), and it received the DHCPOffer message sent by the router (tells you that it is two-way connectivity). This will tell you that it is able to transfer packets, which means that the problem is not in the network itself, but on a device.

If it cannot do that, then the problem is possibly firmware or hardware based. The fact that the system boots up to give you the configuration/display lends one to think that if it is firmware based, it might be a module. The fact that the Link is Up says that the physical port is fine.

You mentioned that you replaced the switch and cable, so that can also be ruled out.

No chance it's not the other devices on the network, or perhaps something either spoofing the MAC address or that IP address is assigned to something else? If the AP20 says it is 192.168.206.x, and either the router or another device says it is 192.168.206.x, you will get problems, possibly malformed packets, as the IP address will bounce back and forth. The moment the other device says it has that address, and might be actively using it, the AP20 might end up being forced out.

Sorry, just thinking out loud.

Leslie

Posted by Christian Hove (Member # 6743) on 12-08-2016, 10:30 AM:

Thanks for all your input!

Long story short: I believe that a temporary IP conflict made the ethernet interface of the AP20 stop responding until power cycled.

I discovered that the DHCP pool of the router was overlapping the IP of the AP20.
At least the timing of the problems correlates with a DHCP client configured device being briefly connected to the network for service purposes.
I made a script pinging all equipment and the AP20s have been answering correctly since I corrected the DHCP server configuration.

Regarding the sybErr, I see sybErr count ranging from 4-7 on all AP20s. Anyone with AP20s that can check for non-zero sybErrs?

Posted by Carsten Kurz (Member # 5396) on 12-08-2016, 12:09 PM:

Should be able to have a look in a few minutes.

- Carsten