<html theme="default-light" iconset="color"><head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head><body text="#485663">Hi Pim,<br>
<br>
<span>Pim van Pelt wrote on 01/07/2025 17:59:</span><br>
<blockquote type="cite"
cite="mid:550c6c73-80f9-4c76-83f9-8155e755455e@ipng.ch">
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
Comparing with just 'snmp walk', I can see IF-MIB::IfDescr pop up
revealing that the Nokia switch is sending strings like these:<br>
iso.3.6.1.2.1.2.2.1.2.16382 = STRING: "ethernet-1/1, \"Infra: csw150
eth-0-3\""<br>
iso.3.6.1.2.1.2.2.1.2.49150 = STRING: "ethernet-1/2, \"Infra: csw150
eth-0-4\""<br>
[...]<br>
The corresponding IF-MIB::IfAlias are:<br>
iso.3.6.1.2.1.31.1.1.1.18.16382 = STRING: "Infra: csw150 eth-0-3"<br>
iso.3.6.1.2.1.31.1.1.1.18.49150 = STRING: "Infra: csw150 eth-0-4"<br>
[...]<br>
And IF-MIB::IfName are:<br>
iso.3.6.1.2.1.31.1.1.1.1.16382 = STRING: "ethernet-1/1"<br>
iso.3.6.1.2.1.31.1.1.1.1.49150 = STRING: "ethernet-1/2"<br>
</blockquote>
<br>
ok, so the standard approach for handling this is to put the interface
name into ifDescr, to put the description into ifAlias and to put the
alias into ifName. Your output confirms that Nokia have taken the
unusual approach of using ifName for interface names, ifAlias for
interface aliases and ifDescr for interface descriptions. I dunno what
they were thinking when they did this, but we'll need to remap these
OIDs in the OSS_SNMP package to make it work.<br>
<br>
I'm guessing that the feature of snmprec that you ran into was at least
one of the reasons why the Librenms people wrote their own snmprec
capture code. Usually it will get to flushing the write buffer to
include the ifDescr OIDs before hanging.<br>
<br>
Nick<br>
<br>
</body></html>