if you want to see what you are advertise, the only really working way is to do a packet sniffing of BGP packets (TCP 179), open it with WireShark, looking for an UPDATE message, then in the body looking for BGP-> NLRI

Ros 7.3.1, dump-saved-advertisements does nothing.. no error, no log, no saved file...
Bgp session with .keep-sent-attributes=yes
routing/bgp/session/dump-saved-advertisements iBGP_IT3-EDGE02-1 save-to=t1.pcap

absolutly notting happens... :-/

Did you find improvements by disabling checksum offload?

I only read about possible problems ..

thanks
Alessandro

Hi,
I have done some tests on the QinQ on Hyper-v hosted CHR, I have tried both the simple Vlan in Vlan and the service tag (QinQ 802.1ad), but I can't get it to work.

has anyone succeeded? I can't found any documentation about QinQ support for Hyper-V.

Thanks
Alessandro

same config on V7b2 is working!! now vlan on brige work as expected also with CHR on Hyper-V! :-)

You have to change switch port until you find the correct sequence that not flap.. good luck!
Mk switch @10gbps are not reliable at the moment

the problem with chr and hyper-v is still here
RouterOs 6.47 on windows 2019, vlan on bridge are lost
I will try on v7 beta..

Alessandro

for you MikroTik = RouterOS ??!?

onestly, yes.

This is user forum. Support replies in some topics occasionally, but there's not guarantee they reply to your particular message. If you are looking for an official reply you should contact support@ and/or you supplier/distributor directly. I think that after one year and an half, many support requ...

Switch to SwOS
Use it with no flapping

ok, you can also buy another vendor switch, but I don't think that it was the question...

Hey Mikrotik guys, where are you?

CRS326-24S+2Q+RM @10gbps have the same problem of crs317.. port flapping, link down.. you have 24 ports but you have to do a lottery moving your link, hoping to found a working solution.. (yesterday: 10Gbps Dac to hpe server nic, on port 5 NO LINK, on port 7 flapping link, on port 9 OK).. very disap...

I changed 2 317 with 2 326-24+
The link with all 328 and between 326-24+ are now fine @10Gbps
One link with 326-24G still flap and needed to be set @1Gpbs
Better then before, but seems to be not fully resolved.
Alessandro

as MK support said: "if possible, please, consider using CRS309-1G-8S+ as replacement in similar applications to decrease probability of encountering this issue."

2019 up up up

new reply by mk support:

We are working on this problem, it could take up to a couple of weeks, to find a proper solution and test it.

we hope it...

Very bad service, client try to use Ubq to replace all CRS328.

Hi gentrice, we are angry too, but the problem we are talking here is on crs317.. what kind of problem do you have on crs328?

Mikrotik support, please WAKE UP!!
Customers are angry and whants different brand hardware!

I can confirm that the problem appear also with 2 CRS317 connected trought mk DAC.
PLEASE MIKROTIK, GIVE US SUPPORT!!!!!!!!!

You have to check it on 328 side, looking for disconnections or rstp recalculation

Mikrotik support tell us that the problem is well known and that they don't have ETA for the resolution.
Their official workarround is to downgrade links to 1Gb or use switchos on 317
I hope that someone take care about this terrible bug!!

Alessandro

After we switched on the CRS317-1G-16S+ to SwitchOS 2.9
5 days without port flapping.

Hi roggles, still stable with SwitchOs?

Me too.. 317 - Ccr 10Gb mk Dac ok 317 - 326 10Gb mk S+85DLC03D ok 317 - 328 10Gb mk Dac flapping The strange thing is that for over one month all worked fine, after that, without any change, the flapping began. Forcing the Dac to 1Gbps stop the flapping immediatly, now we are monitoring it, but we n...

Hi,
are the tx-flow-control and rx-flow-control the mikrotik implementation of IEEE 802.1Qbb?
IEEE 802.1Qbb is a must-have for the RoCE (RDMA over Converged Ethernet) that could be used with 10Gbe on CRS317-1G-16S+RM

Thanks
Alessandro