network disconnecting
Hello at all,
A strange thing happened to me recently with one of my clients. this customer has a geographically large all fiber network with all SG3452X and 10gb sfp. on one of the switches a cable runs to an office. this network cable had the patch cord that passed under the chair and walking on it had destroyed the wires so much that they were touching. this behavior caused all the switches to go haywire and cyclically started to give hearthbeat missed, adoption, disconnect. finding the culprit wasn't easy because we had to isolate switch by switch and figure out which one was affected to figure out which port it was. but is this ever possible? Is there any setting I can set globally to avoid this?
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
Hi @MC20
Thanks for posting in our business forum.
MC20 wrote
the network is vast but the setup is very simple. 8411 router, OC300 and 20 48-port switches connected by fiber to a centeostella 3016; several vlans propagated. spanning tree stp enabled but the network went crazy twice: the first due to a network cable nibbled by the wheel of a chair and the last time by a voip phone that was broken. in both cases all the switches were disconnected and reactivated in rotation, giving me no clear indication of where the problem could be. I had to manually disconnect the bridge doors, starting to isolate the sections until everything calmed down and then working backwards to find the culprit. imagine that on a large network this could take several hours with a production downtime. I expected that I could immediately see the anomaly from the controller. What global settings do you suggest?
OK. I know what you are looking for. So, take a look at the log and share the log with the support team, and learn how to read the logs. There is a running log system on the Omada Controller which reflects the status and connection(of the Omada connected/disconnected/heartbeat missed/adopt).
I am not specialized in the Omada Controller so I don't often read the running logs. It contains mostly about the controller and device status interaction.
So, I think you should find out the running log of the last time you experienced the problem. Send that log to the support and ask for advice on how to read the logs and make use of that.
BTW, STP is useful when you have manually and consciously created the loop to enable the backup link. Or you should configure LAG to enable your network with the failover. I don't know if you actually have them and enable the STP for this reason. So, I don't have a global setting for you or specific suggestions. If you are asking how to find out what's wrong, that running log is what you should check every time when something happens to the controller and the device's status. No better way or global settings are recommended.
- Copy Link
- Report Inappropriate Content
Hi @MC20
Thanks for posting in our business forum.
Check the log if there is any loop.
Replace the wires that wear out.
- Copy Link
- Report Inappropriate Content
yes of course, replacing the cable solves the problem. the problem arises from the fact that the company is large. finding a broken cable that sends me into a tailspin is just luck. Is there a possibility that the system doesn't go haywire but isolates the door where an anomaly comes from?
- Copy Link
- Report Inappropriate Content
Hi @MC20
Thanks for posting in our business forum.
MC20 wrote
yes of course, replacing the cable solves the problem. the problem arises from the fact that the company is large. finding a broken cable that sends me into a tailspin is just luck. Is there a possibility that the system doesn't go haywire but isolates the door where an anomaly comes from?
I don't know your network setup. And what function have you enabled to stop the loop if there is any?
If you want me to give specific instructions, I think I cannot really do that. I need more details about your issue and network design.
Could you tell me what is the reason for this problem? Did it enter the loop? Or just the cable failed. If it is the broken cable, check the LAN interface on the switch should find it out but this is based on a precondition that you have listed in a chart.
If it is a loop, STP/RSTP can stop it and prevent the whole network go haywire.
- Copy Link
- Report Inappropriate Content
the network is vast but the setup is very simple. 8411 router, OC300 and 20 48-port switches connected by fiber to a centeostella 3016; several vlans propagated. spanning tree stp enabled but the network went crazy twice: the first due to a network cable nibbled by the wheel of a chair and the last time by a voip phone that was broken. in both cases all the switches were disconnected and reactivated in rotation, giving me no clear indication of where the problem could be. I had to manually disconnect the bridge doors, starting to isolate the sections until everything calmed down and then working backwards to find the culprit. imagine that on a large network this could take several hours with a production downtime. I expected that I could immediately see the anomaly from the controller. What global settings do you suggest?
- Copy Link
- Report Inappropriate Content
Hi @MC20
Thanks for posting in our business forum.
MC20 wrote
the network is vast but the setup is very simple. 8411 router, OC300 and 20 48-port switches connected by fiber to a centeostella 3016; several vlans propagated. spanning tree stp enabled but the network went crazy twice: the first due to a network cable nibbled by the wheel of a chair and the last time by a voip phone that was broken. in both cases all the switches were disconnected and reactivated in rotation, giving me no clear indication of where the problem could be. I had to manually disconnect the bridge doors, starting to isolate the sections until everything calmed down and then working backwards to find the culprit. imagine that on a large network this could take several hours with a production downtime. I expected that I could immediately see the anomaly from the controller. What global settings do you suggest?
OK. I know what you are looking for. So, take a look at the log and share the log with the support team, and learn how to read the logs. There is a running log system on the Omada Controller which reflects the status and connection(of the Omada connected/disconnected/heartbeat missed/adopt).
I am not specialized in the Omada Controller so I don't often read the running logs. It contains mostly about the controller and device status interaction.
So, I think you should find out the running log of the last time you experienced the problem. Send that log to the support and ask for advice on how to read the logs and make use of that.
BTW, STP is useful when you have manually and consciously created the loop to enable the backup link. Or you should configure LAG to enable your network with the failover. I don't know if you actually have them and enable the STP for this reason. So, I don't have a global setting for you or specific suggestions. If you are asking how to find out what's wrong, that running log is what you should check every time when something happens to the controller and the device's status. No better way or global settings are recommended.
- Copy Link
- Report Inappropriate Content
Information
Helpful: 0
Views: 310
Replies: 5
Voters 0
No one has voted for it yet.