Microtik CCR-1036-12G-4S Weird Activity - Losing Connectivity

One of our PPPoE Router is Microtik CCR-1036-12G-4S, has close to 900-1100 concurrent PPPoE sessions 24x7, passing 500-600Mbps of traffic, with RoS 6.32.4
All of our PPPoE NAS Routers are connected to a huawei s5700 switch which comes from the Core Router CCR1072.


Since the vlan from huawei to 1036 was not working, when we had installed it, we simply let it on vlan1 and kept the WAN IP on a direct interface.
[We had even tried with a bridge and associating the port, like I read that CCR needs a bridge for vlan to work, but nothing worked. ]


Similarly we have 4 other PPPoE NAS Routers connected to the same Huawei, but none of them have any problem. The other 3 are CCR1009 and vlans work very smooth on it with the bridge.

Only this router, is losing connectivity at random hours taking down all the 1000+ customers.
Customers radius log says - FORCE FULLY.
Either its kernel failure or CPU goes crazy or something happens, which I dont know by which the router stays on, but simply loses the internet and hangs and takes down every customer PPPoE sessions and then goes online and connects back everyone.

FYI, my CPU is hardly 15% with all the 1000+ concurrent PPPoE Sessions, so I dont think its a problem with load at all.


In Huawei there is no log about, because the ether never goes down, but till microtik is losing the connectivity to internet.
It simply doesnt respond. Even if I click the front LCD, it shows nothing.

It comes back to the life automatically in 1-2 minutes.
And when I reboot, it hangs at loading kernel and I end up unplugging the cable and plugging back.

So, what would be causing this ?

vlan1 is completely blocked on the access network which goes from this microtik too. So, vlan1 flapping is not an issue.
Interface cable is not an issue, because huawei doesnt log about anything like interface down and up. Interface stays up all the time, but microtik is failing or getting hanged.

Shall i go ahead and ask for a replacement for this immediately ?

In the health, fan says 0RPM for both fans.

Have emailed to microtik support with the support.rif file as well.

Forgot to add :

CCR1009 where the vlans are working are on latest RoS 6.37.2

And this CCR1016 and CCR1036 are on 6.30.2 and 6.32.4

Can RoS be an issue of VLANs not working ?