We just left 2 big Mikrotik swamps, and entered another one
Situation: CRS125-24G-1S, v6.37.3, latest Routerboot firmware (3.33).
Acts as a core router, ports splitted on 3 partitions, first partitions use VLANs etc.
Today it had simply locked up. But it wasn’t regular lockup, It was normally reachable through winbox, all ports was running, everything looks fine, even small part of traffic were switched between ports.
However, only ARP packets was passed, so Machine1 could resolve MAC address of Machine2 w/o any problems, but no any ICMP, UDP or TCP traffic could pass that time.
Soft-rebooting switch (system->reboot) immediatly fixed up problem…
Unfortanently I don’t have supout file, just though reboot will not help in this case…
Does anybody experienced similiar situation? I checked changelogs and no looks like updating to 6.37.5 is pointless for this problem.
CRS have super-stable voltage, over UPS, and use stock power supply (no PoE).
Problem showed up again today,
I have rebooted switch with System->Reboot (so no power-off, power-up), and all back online again.
So it must be SW problem
Has anyone tried swapping in a CRS326 for one of these locking up switches. We have a CRS226 that locks up every few days, and no firmware changes have seemed to help the problem. I think we are just going to have to bite the bullet and get something else. I noticed the 326 has a different chip in it, so maybe it’s worth a try? Just wondering if anyone has tried one yet.
Well, I dont have atm anything else from Mikrotiik to check.
Anyway, what I have on my mind, and didnt had any chance of checking that yey - maybe worth to dump unicast fdb table before reboot, and try to flush fdb ? Maybe ìt would wook up switching chip.
Potentially its possible to make mirroring, but catching packrt of dead (which likely kill switch chip somehow) would be pain in ass.
I have almost 70pcs Crs125 deployed. Problem is observed on core switch (complex conf, ports partitoned in multiple groups).
On others, which have almost default config (+port isolation), i do not observe this problem. Maybe happened once till last 2 years at one site.
Anyway I have opened ticked at 5th of july, however seems MT have no ideas how to move with that.
I had the same issue. I emailed Mikrotik several times, opened a trouble ticket. They had me flash to current rev firmware; it still continued. This is a problem which would take down our entire WISP. I replaced it with a competitor’s model and have had no problems. We are at 100% uptime since that last change (our network is now very reliable).
The problem is that after I sold the MT swtiches, I will NEVER purchase another one.
I love their routers and have not seen the issue on that product line.
If you don’t mind, what vendor product did you use. We need to put something in place of our CRS226 but we need the 2 10Gpbs SFP+ ports. I’m just looking at options now. We have a new 326, but it looks like there might be 10Gbps SFP+ issues on that one right now, so I’m wary.
We switched to the CSS326 and have over 8 days uptime so far. That’s longer than the CRS226 has gone in a long time, so hopefully this new switch will function this well going forward. It’s a really good value for sure if it’s stable.
Well, we just passed the 2 week mark with no issues to speak of. It was almost unheard of for the CRS226 to go 1 week much less 2. Keeping fingers crossed we don’t even have to look at this thing again until we’re ready to upgrade or replace it.
1 month of uptime with no issues so far. I think it’s safe to say at this point that the new switch is not affected by whatever issue was causing the old one to fall over 1 a week or so.
Whether or not that is more related to this one running SwitchOS instead of RouterOS or it’s the different chipset I don’t know. I guess the old CRS226 is going to be relegated to lab duty unless Mikrotik can ever isolate and squash the bug.
