I have a CCR 2004 rebooting time to time. Some times twice in a week, other once per month…
Tried another one with the same issue.
Haven’t tried other model.
FW version is 6.46.8
Log after reboot doesn’t show anything usefull. Seems like if it was a Power Supply issue. (router was rebooted without proper shutdown)
But this CCR has two AC inputs. One of them comes from UPS. And, in the same rack, there are more devices (some with 2 AC inputs and other with only one plugged in UPS) and no one of them reboots. So… seems impossible beeing an AC failure.
We’re in the way of configuring a remote log server to see if there is some log just before CCR reboots.
Has anyone other ideas to solve it?
6.48.2 has specific fixes for stability and packet loss issues with the CCR2004. I’d upgrade to that version and see if it resolves the problem.
What's new in 6.48.2 (2021-Apr-09 10:17):
###clipped###
*) switch - improved resource allocation on 98PX1012 switch chip for CCR2004-1G-12S+2XS device;
*) switch - improved system stability with 98PX1012 switch chip for CCR2004-1G-12S+2XS device;
My fault. Was unable to find it in forum…
Reading right now.
Thanks!
Thankyou.
Well I knew this day will come.
FW version 6.48.x seems to me that is having some kind of VPLS issue. So, we’re stuck in older FW.
I did the check even in emulated environment. But this is another fight that we delayed for too long right now because we focused in another ROS issue on PPPoE with high BW limits… xDD
Didn’t mean it like that. Just wanted to show you that thread in case you hadn’t seen it. Unfortunately, many of us are waiting for the 2004 stability issues to be resolved, and I wanted to make sure you knew your situation isn’t an isolated incident.
Has anyone noticed an improvement with versions 6.49beta22 or higher, or with 6.48.2 and random reboots?
I have several in test with the new RouterOS (we have 25 CCR2004 installed on our network) and I continue to have random reboots (often every 2 to 3 days, sometimes 15 days), with devices that only do OSPF + PPPoE BRAS.
We have about 10-15 CCR2004s in production, several of them running 6.48.1, and they work almost flawlessly with reboots every now and then. We have one core router which rebooted about once a week, but we have been running the betas in our lab without flaws and we decided about two-three weeks ago to upgrade that router to the same beta (big nono) but after testing it in our lab we decided the beta30 was stable enough to be used on that router, and it has been running perfectly since then.
Also, we disabled SNMP, cause that is probably also part of the reboot issue, not sure if it is fixed in the latest betas or not, but Mikrotiks SNMP-implementation is really bad imo, and we can live without it on these routers for now.
VPLS issue involves MTU’s.
I did see it some time ago. Just when I instaled the 2004 at first time. It was a pain in the ass because at first we tried to upgrade every single OSPF/MPS/VPLS router member to match the very exact version.
It’s not for CCR2004. I make a virtual network using EVE that uses x64 ROS images, and the issue is the same. Downgrading to 6.47.x, and it works fine again.
It’s now in our engineer’s hands, that will ask for Mikrotik Support it after testing lattest firmware issue persists
But it’s easy to replicate, actually.
Setup 4 routers: 2, with firmware 6.48, for the OSFP and VPLS (MPLS is optional) and the other 2 for sending pings through VPLS.
Once the ping packet size is big enough it doesn’t reach the remote router, 1474 bytes works, 1476 doesn’t.
No-fragment flag is irrelevant.
Again will create a new thread when engineer tries latest ROS firmware.
For the PPPoE…
We have 6 CCR-1036 running pppoe servers at the same time.
Seems that it’s not a problem just with the amount of pppoe tunnels running simultaneously.
It fails because on the >500Mbps clients when router has more than 200 or 300 clients. Most of them have very low bw profile (<50Mbps) But, it’s almost impossible to reach BW tests > 500 or 600Mbps
