unreachable v bad

[nmitsteve1]

Hi All,

Please can someone explain the differences in Spectrum between a link being "unreachable" and "bad"

Event codes

0x00010d11 The link status of port (name - {m}, type - {t}) is now "bad". This generates a CRITICAL as Cause Code 0x0001040a

0x00010d14 - The link status of port (name - {m}, type - {t}) is now "unreachable" is log only with no associated Cause Code so does not raised an alarm

Help greatly appreciated.

Steve

[STEVE MOORE]

Usually, “unreachable” would indicate that Spectrum has no way of gathering any information on that particular link. Spectrum doesn’t see the polling attempts reaching the link, so it can’t tell what the state might be, good, bad, anything. It has no way of knowing anything about this link.

A “bad” message would indicate that Spectrum has gotten some information about the link, and has determined its state. It is marked “bad” for a reason.

[Michiel Helder]

We usually see a bad link alarm when a connection is not modeled correctly, for example when a server was moved to a different switch port. In that case the server on the other end of the link is still reachable, while one of the ports on the link is down. There might also be other cases where you see this event. My guess for unreachable would be that it indicates a "grey" condition since it is log only. This means Spectrum can't collect data from it, but this is a symptom of an upstream link/device failure so root cause analysis determines it should not result in an alarm. But that's just my best guess without checking any documentation.

[Anon Anon]

I agree with Michael. When you have these alarms the status/condition of the ports will give you more information. Have a look at them when you have these alarms. When it says unreahcable you might also need to check the device reachability or look for any snmp alarms on the device model

kalyan