I also have this router and all is looking normal, dont have any drop packets . I also use 10Gbit sfp+ modules from Finisar.Hallo,
I have a new CCR2004 here and been testing in different constellations.
We only used 10G connection. We connected a Juniper Router, a DWDM System, a Server and a CCR1016.
In any setup we saw significat PL in all directions. The only counter that was running and looked suspicious was the TX Drop on the according outgoing interface. But why are those packets dropped?
We were using hardware-queues only, but the ports did not nearly ran at full capacity and the PL continiously occured anyway. eg. at 100 MBit load
What I consider very sange is that this is the deault setting for a SFP+ Port. This leads to the port links at 10G even though the rate is not selected and behaves ever weirder.
also funny, paket statistics for packets > 1023 missing
Something is very wrong with this router. Don't buy before you hear that this is fixed.
ROS version is 6.47.7, but I think this router is meant for ROS 7 when it will be stable. ARM64 is better in 5.6.x kernel... :)Ros version?I also have this router and all is looking normal, dont have any drop packets . I also use 10Gbit sfp+ modules from Finisar.
ARM64 is better in 5.6.x kernel... :)
I unpack ros7.1 b2 , but looks like there is kernel 5.6.3, maybe i am wrong?ARM64 is better in 5.6.x kernel... :)
I think it was mentioned that ROS v7 will be running kernel 4.19. But I guess this can change before stable v7 gets rolled out, new linux kernel LTS was announced after the rumour circulated (5.4).
I'm sorry if I'm off topic. I can't find any real test results in CCR2004 anywhere. Please tell me how much traffic you have going through the most loaded CCR2004 and what is the load on the CPU ?I took counter samples from three of the seven CCR2004s I have in production and the TX drop rates for the busiest interfaces were:
* 1/23536 packets dropped (0.0042%)
* 1/76566 packets dropped (0.0013%)
* 1/104768 packets dropped (0.00095%)
These are all on 10G ports with 10G fiber links and ROS 6.47.2 or 6.47.3.
Our CCR2004s currently do BGP (full table) and have a simple Raw firewall (conntrack disabled). So with BGP running and no fasttrack/fastpath, we can saturate the main 10G egress interface with web traffic (a few thousand simultaneous flows). One of the CPUs is pegged (the one doing BGP, this is ROS6) and the rest are well-utilized on traffic peaks, but not at 100%. Sorry, don't have more exact figures at the moment.I'm sorry if I'm off topic. I can't find any real test results in CCR2004 anywhere. Please tell me how much traffic you have going through the most loaded CCR2004 and what is the load on the CPU ?
Thats strange.... I use Finisar FTLX8571D3BCL sfp+ modules, test 10Gbit bandwith and no errors. I also use complex firewall with multiple chains and processor is not loaded even 1/3 of 100%.I am checking deep and I have discovered that packet loss is only on the SFP+ port s, using 10 giga module on the SFP28 ports we have no packet loss.
According with the block diagran https://i.mt.lv/cdn/product_files/CCR20 ... 200459.png the SFP+ and SFP28 ports have different bus to the PIPE