Entire site went to inactive bank for SW

We were making some changes and simply unplugged the cable from the AP and plugged it back in maybe 5-15 seconds later (one time).  So far every AP has reverted to the older firmware.  Would this be caused by either the reboot or the inability to receive sync from the CMM?  Or both?

I was under the impression it would only use the backup bank of SW if it failed to boot 10 times.

Note it is configured to receive sync however the CMM is not yet installed.

Hi Josh,

Your understanding is correct.  It should only fallback to the older bank during a repeated failure condition with the current bank.

What version of software were you running when the fallback to the other bank occurred?


We will try to reproduce the problem.

Dan

Josh,

This is very strange issue.  Would you be able to contact our support engineers and provide remote access to them so this can be looked at? When you said site, how many APs on this site fell back to the inactive bank?

Thanks,

Sriram

5/6 of the APs went from 2.3 to the inactive bank which was I believe 1.4.3 (IIRC).  The sixth and final one is giving me link issues to the switch - we're replacing that shortly.

How do I reach the support engineers?

Hi Josh, 

Do you remember if you did any software upgrades to the site just before this happened? The only reason the APs will fallback to the inactive bank is when the active bank is corrupted and it is unable to boot on it. The corruption primarily happens when the radio is writing to flash (ex: during a software upgrade) and power glitches happen. 

Our support engineers can be contacted through phone, email or by opening a support case here: http://www.cambiumnetworks.com/support/contact-support

Thanks,

Sriram

It's a known issue in 2.3.  You have to upgrade twice so that way it fails back to the inactive bank that is also 2.3.

Hi Josh, 

This is not a known issue. I believe the support engineer mispoke when he said it is a known issue. He recommended that you load the same latest software on both banks by upgrading twice. This will ensure that if the APs do fall over to the inactive bank for whatever reason, the y will come up with the same version. 

Thanks,

Sriram

According to support it was a known issue.  You're supposed to upgrade to 2.3 twice which means it will always load up 2.3 in the event their is an issue.

I am installing my first ePMP radios.  I have two of the synced radios setup as a backhaul.  They tested fine on the bench.  Both sides of the link have a CMM3.  When I boot up the AP for the first time, I find that it reverts to the default IP and the old firmware.  It also takes more than 90 seconds to boot and establish an ethernet link.  When I reboot via the gui, it only takes 30 seconds.

Active Bank: 2.3.4

Inactive Bank: 1.0.3

If I upload the new firmware and reboot via the AP gui, it comes up fine and my custom configuration and static IP are still intact.  However, if I power cylcle the AP, it wil boot up with the old firmware again.

Whatever is happing is triggered by a cold start.  Unfortunately, I have not tried to cold start the AP without the CMM3.  It is now deployed on a tower waiting for me to put it into production.

I'm following this advice and doing the upgrade twice to get rid of the 1.0.3 firmware.  It workds.  It still takes 90 seconds to cold boot the device, but it at least comes up to a running state.  However, why would this occur in the first place.  Is there something the AP doesn't like about my CMM3?

Hi Will, 

What is the temperature at the site? ePMP has an onboard heater circuitry to warm up the board before booting up. The heater circuitry turns on to raise the temperature of the board and then, once the temperature is optimal, the board boots up. Depending on how extreme the weather is, boot up may take up to 20 minutes. 

Thanks,

Sriram

Also, is this a 5GHz or 2.4GHz radio?

Thanks,

Sriram

Okay,

I'm having this issue with 4 2.4 GPS Aps running 2.5.1

This morning they were all back to the OLD firmware, one of them is not replying to pings (MAC or IP), and NONE OF THEM have subscribers connected to them. It's the first really cold day we have had, but Ive been having this problem of unresponsive radios during warm weather.

I cannot continue going on this way.

We have one 5.8 AP doing this. Had clients dropping session for no clear reason, decided to swap out the AP radio. (no change, btw) Was running 2.6 w 2.0 or something (later vers still dark bg) in secondary bank.

Plugged up on bench, it boots the old fw, and doesn’t show but one flashbank in fw page. (don’t remember if those old fws did that)

Upgraded (again) to 2.6 and looked good. Guys mounted it on a tower again today, and back to same old fw. (reflashed again, both banks this time)

j

i got the same problem a week ago. First put a 802.3af POE on it its help i see glitch when use long cable with the 750UP POE and original POE from cambium. Second update your firmware to 2.6.1_RC1.  here the link

https://goo.gl/4TpMif

use SMPX for NON-GPS and APPX for GPS Sync radios.

and from there the problem was fix :)

We're using a 5.8 ePMP radio to backhaul two towers. After a power outage we had a problem with one of them continuously defaulting from 2.6 to 2.3.4. At which point we would have a myriad of problems. 

What seems to have finally fixed the issue was uploading 2.6, and then uploading it again so that it is in both the active, and the inactive bank. Since I've done this (8 days ago) we haven't had any problems with the radio. Where as prior to that it wouldn't run more than 8 hours without defualting to it's inactive bank.

I am wondering if somehow during our power outage if the 2.3.4 firmware was somehow corrupted and was  the culprit to the issue. 

1 Like