Strange 60seconds packet loss

I have three sites, each of them has an ADSL isp router with its “LAN” ethernet connected to RB2011UiAS “WAN” ethernet
All ports and protocols are transparently forwarded from router dsl interface static pubblic ip address to routerboard “WAN” ipaddress.

I have opened two terminals each routerboard to ping continuously the other two public ip addresses.

Sometimes , between system A and B, or A and C or B and C, without a logical cause, I experience a contiguous lost of 60 packets (timeout), then pings came up to life again.

There is no lost of dsl line or connectivity, in fact , from a fourth site , “pinging” continuously all three sites public IP I have NEVER any packet loss.

Each of three sites has a static route :

A S aaa.bbb.ccc.0/24 (public class C ip address to which ISP routers belong) 10.0.0.1 (lan of ISP router) 1

What should I check to understand the problem ??

Each of three sites has a static route :

A S aaa.bbb.ccc.0/24 (public class C ip address to which ISP routers belong) 10.0.0.1 (lan of ISP router) 1

Have you tried without this static route?

Is it possible to put those DSL routers in bridge mode?

I’ve forgotten to mention the static route is because this connectivity is not the main of routerboards system
the default route 0.0.0.0/0 is routed to a different (internet) router attached to eth2 on each routerboard.

I’m waiting for ISP to check for interruction from inside their routers , but something tells me there are some RB issues…

As I’ve said a foreign iternet connection pinging routers public ip addresses has never disconnections, no packet losses neither if I ping continuously ISP router LAN from RB eth…

It isn’t possible to put routers in bridge mode (are you thinking to a portforwarding/nat issue on the isp router side ? )

You’d better post a complete export (edit out any sensitive information).

Are you using ECMP?

Is ROS and firmware up to date?