Old thread, but I use the CRS317 with 10Gb SFP+ modules in LACP 802.3ad bonded to the servers which run Intel X520-DA2 NIC’s with 2 ports as trunk ports to handle the tagging. I use both XCP-ng and Proxmox clusters for the hypervisor functions. However I use an rb5009 for all routing, trunked to the CRS317 - so my setup is a bit different in that I don’t use pfsense. I run RouterOS on both devices, version 7.16, all runs fine for my needs. I host about a dozen cloud workloads over it, and I need the 10G links for the shared storage VLAN mostly. I run the corosync network over a standard 1GB uplink.
Best of luck on your setup.
Dear Team,
I am facing some issue with CCR-1072
loading kernel… OK
setting up elf image… OK
jumping to kernel code
(0,0) hv_panic: msh3: detected uncorrectable memory error on rank 0 (H)
System halted.
(6,1) hv_panic: got read error response on RDN interrupt: PC 0xffff_fffc_0003_8e88, ICS/PL 0x3
Looks like it is ready for the trash-bin then.
The CCR1072 has 16GB embedded (1-bit error correcting memory) ECC memory, so if the kernel sees an “uncorrectable error” I think it really busted
As far as I know you cannot just replace memory-modules.