WS-12-250-DC Segmentation fault

[darimar]

Hello,
today I've installed a brand new WS-12-250-DC. It worked fine for about two hours and after, without any reasons, It was unreacheable. I pulled off it from the production site and I tried to test it in my lab. No response from web ui, and only a continous response from serial console: "segmentation fault".
What I have to do in this case? I have several type of Netonix in my network, but I never experienced this issue.

Thanks and Best regards,
Dario Martino.

[sirhc]

Please do the following to get a baseline starting point.

1) Hook up your console cable
2) Have someone hold the default button in for 30 seconds when you apply power to the device.
3) You should see a message on the console that that the default button was pressed?
4) Wait several minutes as this process takes awhile to reformat the flash partition.
5) After it boots see if the device allows you into it, if it does please upgrade to v1.4.7rc7

Let me know if you are able to do this successfully.

[darimar]

Hello Chris,
thank you for your fast response. Yes, I completed the procedure. After the reset I was able to access in web gui and from here I upgraded to 1.4.7rc7 (from 1.4.6)

It seems to work normally. Do you think I can put it in a production environment ?

Best Regards,
Dario M.

[sirhc]

I would assume so, I would play with it on your bench for awhile but yes.

Did you import the configuration last time or manually enter it in?

If you imported the config then do not do that as the config you imported for some reason was corrupted or from a previous board Rev maybe?

RevF boards have differnt POE options per port than Rev E and below.

Also if you sparked the DC connection you can cause the radio to corrupt the flash and or load improperly. We do recommend a DC BREAKER/SWITCH but if you connect cable make sure the connection is FAST and solid with as less sparks as possible.

[darimar]

Ok Chris,
now I know what is the problem. I've imported the configuration from a WS-12-250AC rev E. The revision board of the WS-12-250-DC is F.

Now, I enter the setup manually, I try it at the bench for a while...And after I try It in production.


Thank you very much for your support!

Best Regards,
Dario M.

[sirhc]

Yea we recently discovered this BUG with the Port POE options changing and there is a fix in the latest v1.4.7rc7 that SHOULD account for incorrect POE options on a port in a config being imported but I am sure this will be a source of problems for awhile.

Since you have this unit in your office you might want to consider these MODs "if" you are a GOOD solder. see below:

The MODs below bond DC negative to Earth Ground which will help protect the switch from Ground potential shifts between the tower and the AC Service Grounds. We decided to do this because so many people do not do grounding properly but even if they do this is recommended and has been the standard since September 2016.

If the MFG date is after September 9th 2016 then it was done.

You can easily test this with an Ohm meter and measure resistance between PIN 5 of the console port and the ESD shielding on the RJ45 stack if you have a Rev F board or newer. It should read near 0 Ohms meaning the MOD was done.

If a Rev E or older board and a DC powered switch you can measure Ohms between the DC terminal and the ESD shielding which should yield close to 0 Ohms which means the MOD was done.

If you have a Rev E or older units MFG before 9/9/2016 then it is harder to check for and would require the chassis to be opened and inspected.

If you check with your local electronics supplier you should be able to purchase a 0 Ohm Resistor or Jumper Block to solder in place, should be at least capable of 10 Amps. Another option is to solder piece of copper or copper wire between the pads.

THIS SHOULD ONLY BE ATTEMPTED IF YOU ARE WILLING TO BREAK WARRANTY SEAL VOIDING YOUR WARRANTY AND YOU ARE CAPABLE/QUALIFIED OF PERFORMING SUCH MODIFICATIONS. WE ARE NOT RESPONSIBLE IF YOU DESTROY YOUR SWITCH.

This is the 0 Ohm Resistor we use and suggest:
Digikey part: http://www.digikey.com/product-detail/e ... ND/5118869

O Ohm.png (10.81 KiB) Viewed 9056 times

CLICK IMAGES BELOW TO VIEW FULL SIZE
WS-12-250-XX Rev A-E

WS-12-250-XX Rev F

WS-6-MINI Rev A-E

WS-8-150-DC Rev F

WS-24-400A/B

[darimar]

Thank you Chris,
fortunately I'm a good solderer. Tomorrow, I'll do the check for grounding and If necessary I'll do the mod as you explained. It's a very easy and fast operation.

I'm sure that I've some devices manufactured before Sep 2016...But, If the tower has the same ground of AC service and the switch is well grounded, do you think that the mod could give some benefits ?


Best Regards,
Dario M.

[sirhc]

Thank you Chris,
fortunately I'm a good solderer. Tomorrow, I'll do the check for grounding and If necessary I'll do the mod as you explained. It's a very easy and fast operation.

I'm sure that I've some devices manufactured before Sep 2016...But, If the tower has the same ground of AC service and the switch is well grounded, do you think that the mod could give some benefits ?


Best Regards,
Dario M.

Yes even if the tower ground and service grounds are bonded good I PERSONALLY WOULD STILL DO THE MODs.

To give you an idea, since implementing this small change (September 2016) we have not had a single RMA that was from ground current damage to units with the MOD but still see damage on units without the MODs.

Problem is most people do not run dedicated grounds up the towers and to each piece of equipment and make sure the ground path is shorter (this is where Ethernet Service Sloops come in play to lengthen the cable path) and less resistant than the Ethernet cable. But even if they do this will also help against ESD and other types of current damage.

I would do it, and I am doing it at my WISP, rotating units to apply the MODs to then applying the MODs to those I recover....rinse and repeat until all units have MODs.
I should be done before spring rains/storms.

[sirhc]

I will post up more pictures showing the MODs to all other models tomorrow.

[wirelessblue]

I have had this same Segmentation fault problem on two units. One is a WS-12-250A that happened several weeks ago and one is a WS-8-250-AC that happened at 4:30am this morning after I changed a Flow Control setting on a WS-6-MINI attached to it. Both are more than a year old and I now have both on my bench. After the hard reset (30 second...), they come back normally. Can this modification be done via RMA? I am happy to pay for it if need be as well.