All committers have signed the CLA.

Testing this on my setup at the moment and so far so good... Some messed up data seems to be fixed now on newer software versions.

In your issue you say "When running this on a Juniper EX3300 (at least seen on this types) the neighbor info is incorrect most of the time."

What causes it to be incorrect most, but not all of the time?

The device outputs neighbors as ifindex

I get that but you make it sound like it doesn't happen all the time.

I get that but you make it sound like it doesn't happen all the time.

Some ports had the correct data while others did not. There must be something wrong with the check

} elseif (isset($dot1d_array) && isset($dot1d_array[$entry_key]) && is_numeric($dot1d_array[$entry_key]['dot1dBasePortIfIndex'])) { $ifIndex = $dot1d_array[$entry_key]['dot1dBasePortIfIndex']; }

... as the default is to use the array key as ifindex as explicitly done with junos on this commit, which makes me think this is related to #18145

I'm not sure, but I think this is because the statement;

        } elseif (isset($dot1d_array) && isset($dot1d_array[$entry_key]) && is_numeric($dot1d_array[$entry_key]['dot1dBasePortIfIndex'])) {
            $ifIndex = $dot1d_array[$entry_key]['dot1dBasePortIfIndex'];

is just before

        } else {
            $ifIndex = $entry_key;

and maybe not in all cases the elseif is AND statements are all matched and therefor the $ifIndex = $entry_key is also used as in the latest else statement.

An example of data is shown below.

If I add the two lines so $ifIndex = $entry_key is used the LLDP info for this port is correct. otherwise the LLDP info shows local port ge-0/0/27 (which is even down) for the LLDP neighbor.

SNMP[�[0;36m'/usr/bin/snmpbulkwalk' '-Cr10' '-v2c' '-c' 'COMMUNITY' '-OQUs' '-m' 'IF-MIB' '-M' '/opt/librenms/mibs:/opt/librenms/mibs/junos' '-t' '10' '-r' '2' 'udp:HOSTNAME:161' 'ifDescr'�[0m]

ifDescr.517 = mge-0/0/1
ifDescr.543 = ge-0/0/27
ifDescr.552 = mge-0/0/1.0
ifDescr.591 = ge-0/0/27.0
SNMP[�[0;36m'/usr/bin/snmpbulkwalk' '-Cr10' '-v2c' '-c' 'COMMUNITY' '-OQUs' '-m' 'IF-MIB' '-M' '/opt/librenms/mibs:/opt/librenms/mibs/junos' '-t' '10' '-r' '2' 'udp:HOSTNAME:161' 'ifName'�[0m]

ifName.517 = mge-0/0/1
ifName.543 = ge-0/0/27
ifName.552 = mge-0/0/1.0
ifName.591 = ge-0/0/27.0


SNMP[�[0;36m'/usr/bin/snmpbulkwalk' '-Cr10' '-v2c' '-c' 'COMMUNITY' '-OQUs' '-m' 'IF-MIB' '-M' '/opt/librenms/mibs:/opt/librenms/mibs/junos' '-t' '10' '-r' '2' 'udp:HOSTNAME:161' 'ifOperStatus'�[0m]

ifOperStatus.517 = up
ifOperStatus.543 = down
ifOperStatus.552 = up
ifOperStatus.591 = lowerLayerDown

  517 => 
  array (
    'ifDescr' => 'mge-0/0/1',
    'ifName' => 'mge-0/0/1',
    'ifAlias' => 'description of the interface is removed for privacy reasons',
    'ifType' => 'ethernetCsmacd',
    'ifOperStatus' => 'up',
  ),
  543 => 
  array (
    'ifDescr' => 'ge-0/0/27',
    'ifName' => 'ge-0/0/27',
    'ifAlias' => '',
    'ifType' => 'ethernetCsmacd',
    'ifOperStatus' => 'down',
  ),
  552 => 
  array (
    'ifDescr' => 'mge-0/0/1.0',
    'ifName' => 'mge-0/0/1.0',
    'ifAlias' => '',
    'ifType' => 'propVirtual',
    'ifOperStatus' => 'up',
  ),
  591 => 
  array (
    'ifDescr' => 'ge-0/0/27.0',
    'ifName' => 'ge-0/0/27.0',
    'ifAlias' => '',
    'ifType' => 'propVirtual',
    'ifOperStatus' => 'lowerLayerDown',
  ),


SNMP[�[0;36m'/usr/bin/snmpbulkwalk' '-Cr10' '-v2c' '-c' 'COMMUNITY' '-OQUsetX' '-m' 'BRIDGE-MIB' '-M' '/opt/librenms/mibs:/opt/librenms/mibs/junos' '-t' '10' '-r' '2' 'udp:HOSTNAME:161' 'dot1dBasePortIfIndex'�[0m]

dot1dBasePortIfIndex[517] = 543

Is this on a newer software and hardware?

Is this on a newer software and hardware?

Yes, the specific output in #18318 (comment) is from and EX4100-48MP running 22.4R3-S4.4

SNMP['/usr/bin/snmpget' '-v2c' '-c' 'COMMUNITY' '-OQUs' '-M' '/opt/librenms/mibs:/opt/librenms/mibs/junos' '-t' '10' '-r' '2' 'udp:HOSTNAME:161' 'JUNIPER-MIB::jnxBoxDescr.0' 'JUNIPER-MIB::jnxBoxSerialNo.0' 'JUNIPER-VIRTUALCHASSIS-MIB::jnxVirtualChassisMemberSWVersion.0' 'HOST-RESOURCES-MIB::hrSWInstalledName.1' 'HOST-RESOURCES-MIB::hrSWInstalledName.2']

jnxBoxDescr.0 = Juniper EX4100-48MP Switch
jnxBoxSerialNo.0 = <removed>
jnxVirtualChassisMemberSWVersion.0 = 22.4R3-S4.4
hrSWInstalledName.1 = No Such Instance currently exists at this OID
hrSWInstalledName.2 = "JUNOS Base OS Software Suite [22.4R3-S4.4]"  

Here's the deal. LLDP supports different TLV. The code tries to guess what TLV is available by first checking if the 802.1D index is used. Then it should fallback to ifIndex. (there are signs that check is faulty)

These things are sometimes user-configurable and some are determined by the neighbor device. So setting it per-os is usually folly. There are SNMP OIDs to detect which TLV are used, but they are often inconsistent.

Over all, the LLDP code is just a mess and devolving rapidly. Hopefully, I can get some time to push @Npeca75 's rework of it across the finish line.

In summary, you can't just assume it is ifIndex for all JunOS.

Is this on a newer software and hardware?

Yes, the specific output in #18318 (comment) is from and EX4100-48MP running 22.4R3-S4.4

SNMP['/usr/bin/snmpget' '-v2c' '-c' 'COMMUNITY' '-OQUs' '-M' '/opt/librenms/mibs:/opt/librenms/mibs/junos' '-t' '10' '-r' '2' 'udp:HOSTNAME:161' 'JUNIPER-MIB::jnxBoxDescr.0' 'JUNIPER-MIB::jnxBoxSerialNo.0' 'JUNIPER-VIRTUALCHASSIS-MIB::jnxVirtualChassisMemberSWVersion.0' 'HOST-RESOURCES-MIB::hrSWInstalledName.1' 'HOST-RESOURCES-MIB::hrSWInstalledName.2']

jnxBoxDescr.0 = Juniper EX4100-48MP Switch
jnxBoxSerialNo.0 = <removed>
jnxVirtualChassisMemberSWVersion.0 = 22.4R3-S4.4
hrSWInstalledName.1 = No Such Instance currently exists at this OID
hrSWInstalledName.2 = "JUNOS Base OS Software Suite [22.4R3-S4.4]"  

Thanks. This explains why I never saw this on older devices like EX2200.