Strange behavior MINI

[Stephen]

Stephen, strange thing in the access control panel with this firmware. It mentions:
The configured access controls do not include the IP address you are currently using (::ffff:46.**.***.**).
If you save now you will lose access to the device.
I logged out of the switch after upgrade and logged back in and the message is still there.

So what is happening here is the UI is comparing the IP addresses in the access list to the IP address it thinks the UI is opened up with. I've done a couple tests and it appears that even with the right IP it won't match it exactly because PHP is prepending "::ffff:" to the result. Basically, ignore it, as long as your IP address of the computer you are accessing the switch from is in the list, you will still be able to access it.

Also this UI bug will be fixed on the next release.

[Stephen]

Jma,

1.5.1rc6 is out.

It should fix the issue explained above in the Access Control page.

viewtopic.php?f=17&t=240

[jma]

@Stephen,

We will update the switch. In the meantime we have found out that the issue is not resolved. On the contrary, we now have it approx. 10 times a day.
We will change all cables and power supply on the switch again as we really do not know anymore what to do.

[jma]

We have now replaced the switch again. Installed latest 1.5.1rc6. We have also replaced all LAN cables for all the devices that are connected to the switch. Also changed supply adapter and power supply cord. Even put the power cord in a different power supply. Everything went fine for about 7 hours and we even thought it would be solved. And then again the same problem. Switch did not respond for 20 minutes and puts everything off line doing so. Then suddenly reboots and syslog just mentions cold boot.
So we have this problem with all 3 MINI switches that we have tried. It seems useless to keep on trying other MINI switches. As we only have spare MINI switches I can not try an 8 or 12 port cooled switch at this moment.
We are certainly living in a very warm country, so I have now put a cooling fan on it, although we never had this problem in the past years and temperatures were the same.
Anyway, this has dropped the temp from Board 59°, CPU 93°, PHY 93° to Board 44°, CPU 73°, PHY 66°.
Hopefully it has something to do with the heat and this solves the problem.

[sirhc]

This is not a temperature issue.

Did you change anything else recently such as change devices from 100M to 1G links or add more devices on the switch?

[jma]

No, we have not changed anything. The switch is working fine since we have put the fan on it. It is now over 13 hours, which is already a success. Of course I understand that this could be not related.

[jma]

It is more than 36 hours now since we have put the fan and not a single interruption. It looks like this has solved our problem.

[sirhc]

It is more than 36 hours now since we have put the fan and not a single interruption. It looks like this has solved our problem.

The problem I have with this is we have thousands of these units out there working in very warm environments.

We have tested these units to 50C environment with no issue.

I personally have one that is running in a box on a ceiling in a metal shop that makes the chassis that gets above 110 F in the summer so the ceiling temp would be much warmer and it is in a small plastic box (the one we sell) and no issues.

[sirhc]

Can you try this test as outlined in our Bench Test thread:
viewtopic.php?f=6&t=2780#p19221

If you have a WS-6-MINI or WS-8-150-AC (manufactured over a year ago) and you think it is rebooting and or locking up due to the 3.3V CAP issue I would perform this check as well:
Set the switch on bench and create a 1G Ethernet link on each port.
You can use a POE device or a non POE device, makes no difference.
Let the switch run over night and make sure no reboots occurred by checking the up time on Device/Status TAB

[jma]

Hi Chris, thanks for your reply.
The environment temperature is between 40 and 50 degrees in these months of the year.
I will do the test as soon as possible.