[BUG] XG v16/17 PPPOE doesn't endlessly tries to reconnect

Thanks emmosophos 

i just answer now to his request, and i am very glad that a patch seems to be ready for testing !

i will follow up asap.

Regarding 18.0.2, i can only warn you that it is totally bloated. We tried it between two of our sites and we had to downgrade them to 17.5mr14 as we have many vpn and the disconnection of one of them cutted down all the NAT/DNAT sessions.

Hello Guillaume,

Thank you for the follow-up.

I will also be following the case, and please update the thread after the patch has been applied.

Regards,

I now have another customer who is having this issue. They run a XG135 with V18.0.1 MR1. 

The Sophos unit has worked perfectly for 3 years now, until about 1 month ago.

This unit is doing PPoE through a 50Mbps Fibre service.

The unit is in a COVID lockdown in Melbourne Australia, and all the staff are working from home through this unit via SSL VPN. 

There is no one in the office to restart it, so I have to rush into the city to press the CONNECT button. This has happened twice in 24 hours.

The unit has a 4G WWAN USB dongle on it, exactly for this type of issue. However, when this fault occurs the 4G link is NOT started up.

I don't have any issues with other Sophos unit on ADSL or EFM connections with PPoE. I suspect that the issue relates when the PPoE link quickly goes down and up. (1-2 seconds). The PPoE will terminate, but the Sophos unit thinks that the exiting PPoE session is still going. With the slower technologies such as ADSL the link speed is slow, and so the PPoE service does notice the session dying.

I have open Sophos cases 03041322 and 03041326

emmosophos 

Hi, no news of sophos support since 2 days.... seems than this problem is definitly not handled seriously.

Ridiculous.

If you love bad jokes open a case for this issue: they told me it’s under investigation but it might finally be classified as a normal behavior.

three of my units suffered this issue once this week and another one almost everyday.

Sophos I love you!

Hello Guillaume,

Thank you for the follow-up.

You should be receiving an update today from the engineer, as your case has been escalate to get the patch applied.

Regards,

This is a concerning development.

I did see this issue once only on a V18.0.0 unit which I was using for testing, but since upgrading that unit to MR1 396, it has not exhibited the issue again.

Gavin,

My failure this week was with a customer running SFOS 18.0.1 MR-1-Build396.

So it isn't fixed in this release!

So the randomness of this issue continues.

Last night there was carrier maintenance on a fiber service. The attached XG210 unit had exhibited the PPPOE fault once only.

Link was down at 2:19am

PADI packets every 5 seconds for 1 minute then a PAD0 timeout and a session restart

It continued this for 55 minutes while the service was offline then re established connection correctly.

This unit is still running 17.5.MR8 firmware.

Andrew,

That is why it is concerning.

I just got 2 XG125 Rev3 devices which I plan to use as interim staging devices for upgrades to V18. Essentially load their current configuration onto the XG125, default the customers unit and upgrade to V18.0.MR1 and then rewrite the config, as I have experienced issues with the migration on the couple of units I have tested this on, and have ended up reverting to V17.5

Even the upgrade from 18.0.0 to MR1 has butchered the NAT rules attaching some 13 digit number to the descriptions of them all, and completely screwing the name of 1, or it generated a new rule with some self generated name.

BUT SINCE THERE IS NOT A CONFIGURATION DOCUMENTATION TOOL AVAILABLE FOR SOPHOS XG DEVICES! you either have to screenshot every entry and every change, so you can see what has been done.