WS-12-250-AC potential mangled/truncated packets?

[mpeenie]

Howdy,

We have a site which uses a Mikrotik, a WS-12-250-AC, and two back hauls with OSPF sessions to each adjacent site. We had a power outage a while ago and lost one of those OSPF sessions. We have tried everything from upgrading/downgrading the radios, upgrading/downgrading the switch, new ports on the switch, rebooting all gear on site, etc. No settings have changed at the main site, or the adjacent sites. No ICMP packets or small packet sizes could be sent via the link. We could see local ARP on each switch port, but it seemed as if the switch wouldn't allow the ports to speak with each other. Traffic seemed to trickle through, allowing OSPF to INIT, but the neighbor was stuck in exstart(Juniper) as if it was an MTU issue; We verified MTU throughout the link, with no luck. It seems like the traffic was being mangled or truncated for some reason. This link was previously active for over a year with no issues, until this power outage.

Traffic is still flowing properly through the other ports out our other OSPF link, all while we were troubleshooting the main issue.

We put a bandaid on the issue by bypassing the switch altogether and plugging straight into the Mikrotik. Since we've done that we have had no issues with the link. Any ideas on what could be causing this issue?

Thanks

[sirhc]

Were you using QinQ or Double Tagging in the VLAN config?
QinQ and Double Tag got broken in v1.5.12, you could use v1.5.11 so long as it is not an WS-8-150-DC which it is not

Things I would do.

Bring the switch back to your shop and test it out:
Factory default it.
Use 2 laptops and transfer data between laptops and make sure your not getting any errors and getting 1G links on all ports. you can view the port stats with the little gear on the far right of each port row.

Verify the all POE options are working
I would use an AFX radio if you have one as it supports all 4 POE options 24V/24VH/48V/48VH

First Turn off all POE and make sure no port attempts to power up the AFX, this verifies no MOSFETs are damaged and powerd on all times.

Then turn on following POE options and check them out.
24VH check ports 1-2
48VH check ports 1-4
24V check all ports
48V check all ports

Verify on Status Tab all current sensors are working properly while you test each ports POR, when using VH if you hover over watts it should break it down to 2 readings as VH uses 2 current sensors.

Lastly verify your SFP modules.

Again I also suggest linking to the SFP port and warm reboot it a couple times to make sure the SFP module is linking back up after warm reboots. You should also verify by power cycle reboot a couple times.

If all these test pass then the switch should be working fine.


Although if you have a spare you could replace the unit with your spare and see if problem goes away, if problem persists and this site worked for over a year then well who knows.

[mpeenie]

Figured I'd bump this thread since it's been a while since we've dealt with this switch LOL

We upgrade/downgraded the switch that was on site, and ended up swapping with a brand new one. We still were getting the same issue, so the BH is still bypassing the switch into the Mikrotik that is on site. Haven't had to dig into this site since we first experienced these issues considering it's worked fine plugged straight into the tik (Minus some power outages from large storms, but that's unrelated).

I think this site is just cursed LOL We plan on moving this problem site to Juniper switching/routing gear, so hopefully we don't see similar issues at other sites.