I am going to comment on a case that has happened to me and that I have reported but they indicate to me that it has no solution, and I honestly do not believe that the community can be functioning that way.

Currently I have a CCR1072 with 1 Full BGP 10Gb and 70 sessions Peering in a IX.

With all Peers UPs, I have approximately 810,000 routes.

If when it is all working I try to see which ad routes (I have a previous post due to problems in the advertisement of the routes) with "/ routing bgp advertisement print" and no peer filter, the system becomes unstable, so I decide to cut with CTRL + C, at that time the CCR1072 closes peer routes.

If I perform the same operation but specify the peer with "/ routing bgp for print PEER_X", the system does not fail and shows me my routes.

If you filter or not the system you have to consult all the routes, what is the reason that without filter fails and with filter no?

This is the "solution" proposed by support, but neither I see it logical and much less functional in this type of systems:

Unfortunately there is no fix for this problem in ROSv6. You will need to use other means to check advertised messages, for example sniffing outgoing packets.

Do you work sniffing packages instead of using the commands that the system gives you?

I wonder if a CHR has the same issue? I have been hesitant of the 1072, it sounds cool - 72 cores except BGP uses one of those and I believe they are all 1 GHz CPUs.

That said, printing a database table should never seg-fault regardless. Worst case is output should be slow. I'm with you, the "sniff packets" workaround is a joke. 1) Fire developer. 2) Hire new developer 3) Code better or fire that one too.

You can see what kind of networks are advertised. It is impossible to work with such a system.

I Find a workarround:

THIS CAUSES THE FAILURE:

/routing bgp advertisements print where name="PEER1"
/routing bgp advertisements print where name ~ "PEER"

THIS WORK:

/routing bgp advertisements print PEER1

This shows that this is an implementation problem on the part of Mikrotik and that it can be corrected, because if it works in a way that is supposed to do the same as the other and in the other fault shows that something does not work as it should.