XI Network Wizard - Bandwidth speed issue

[lgroschen]

Got it! I will get back to you when I get a version of this ready for you to test.

[lgroschen]

I fixed the warning and critical percentage calculations to retain two decimal places so they should be much more accurate.

Now, can you run the cfgmaker command from the command line?

Code: Select all

/usr/bin/cfgmaker --show-op-down --noreversedns --zero-speed 100000000 --enablesnmpv3 --snmp-options=:::::3 --username=user_name --authprotocol=md5 --authpassword=pass_word --contextengineid=0 10.128.251.45

The above command is what you sent with your mrtg.cfg file so you can use this as I just need the output of the ifHighSpeed. You can add |grep ifHighSpeed to trim the output of the command

Also, like in the page you posted before can you run this command on your switch and see what it's reporting for the speeds:

Code: Select all

show snmp mib get ifHighSpeed

===========
Edit:

So after looking at some other code dealing with this issue they handle it specifically for the Cicso NX-OS, which you are using, and the "nominal" speed of 3705032704 is actually the max speed that can be read for 8GB interfaces. So in the following example they are trying to find out:
1. If speed >= 4294967294 max to use HighSpeed
2. If speed == 3705032704 AND it's from a cisco node (device) force use of HighSpeed

Code: Select all

if ($IfSpeed and $IfSpeed ne ""){
                if ($IfSpeed >= 4294967294) {
                    # Too high for this counter (cf IF-MIB)
                    # get ifHighSpeed table (units of 1,000,000 bits per second)
                    if (defined $refhSNMPResultIfHighSpeed->{"$Index"} and $refhSNMPResultIfHighSpeed->{"$Index"} ne "" and $refhSNMPResultIfHighSpeed->{"$Index"} != 0) {
                        $IfSpeed = $refhSNMPResultIfHighSpeed->{"$Index"} * 1000000;
                    } else {
                        logger(1, " $ifName($Index) -> interface speed exceeding standard counter ($oid_ifSpeed{'oid'}.$Index: $IfSpeed). ".
                            "No high speed counter available, so the speed should be incorrect.");
                        $grefhCurrent->{MD}->{If}->{$ifName}->{MsgCrit} .= "<div class=\"critical\">".denormalize($ifName)." ($Index): interface speed exceeding standard counter. No high speed counter available, so the speed should be incorrect.</div>";
                    }
                } elsif ($IfSpeed == 3705032704 and $ghOptions{'nodetype'} eq 'cisco') {
                    # At least for a problem with Cisco NX-OS, which has an ifSpeed limit/max of 3705032704 for some 8G interfaces
                    # Forcing to take the ifHighSpeed instead
                    logger(1, " $ifName($Index) -> interface speed matching a situation where ifSpeed limit is 3705032704. Forcing the use of ifHighSpeed instead.");
                    if (defined $refhSNMPResultIfHighSpeed->{"$Index"} and $refhSNMPResultIfHighSpeed->{"$Index"} ne "" and $refhSNMPResultIfHighSpeed->{"$Index"} != 0) {
                        $IfSpeed = $refhSNMPResultIfHighSpeed->{"$Index"} * 1000000;
                    } else {
                        logger(1, " $ifName($Index) -> interface speed exceeding standard counter ($oid_ifSpeed{'oid'}.$Index: $IfSpeed). ".
                            "No high speed counter available, so the speed should be incorrect.");
                        $grefhCurrent->{MD}->{If}->{$ifName}->{MsgCrit} .= "<div class=\"critical\">".denormalize($ifName)." ($Index): interface speed exceeding standard counter. No high speed counter available, so the speed should be incorrect.</div>";
                    }
                }

(credit for above code goes to the developer of interfacetable_v3)

This is where cfgmaker would need to be updated or patched to deal with this specific issue. I can maybe put together an update for you to test on your system and then propose a change to cfgmaker through the creators.

[brdr]

Apologies for not getting back to you quicker.

1)
When upi adjusted warn and crit percentages to retain decimal do the calculations now reflect warn at 75% and crit at 90% (as example) properly? If I remember correctly this was an issue as well.

2)
Attached is high speed output snippet. Note: can't grep for ifHighSpeed, and or use --output=out.txt as does not capture what you need.

3)
for some reason the 'show' command that you asked for on these MDS switches doesn't work. Let me see if I can figure this out.

It is interesting that there is specific code for NX-OS. We have about 200 cisco devices, in which, a handful are Nexus OS. All throughout the day when Nagios checks these NX devices we get approx 2600 SERVICE ALERTS showing up in Nagios log for a simple port Status check.

The issue is that Nagios does a check and gets/records a 'UNKNOWN' for the first port 'Status' attempt. Then the second attempt always succeeds. It is ONLY the Nexus-OS devices we see this behaviour!!!

Here are 2 examples (seen within nagios log)...
[1434981993] SERVICE ALERT: as1.nxdev;port-channel154 Status;UNKNOWN;SOFT;1;UNKNOWN - No info is being retrieved.
[1434982012] SERVICE ALERT: as5.nxdev;vpc Peer link Status;UNKNOWN;SOFT;1;UNKNOWN - No info is being retrieved.

[lgroschen]

I have posted a request for the update of cfgmaker, but to be totally honest I'm not sure if it will be added since I haven't heard even a peep from the developers.

The new percentages should be good since they use 2 decimal points now.

If you could find a command that is similar and will run on the device it would definitely help. There may need to be some fixes for the wizard that will deal with the Nexus OS, but for now I'd just like to gather information as to why those devices would be different.

[brdr]

Hello,

Is there any update regarding cfgmaker?

Thx

[lgroschen]

None, but I think we might get a fix together ourselves, test it and then give to you for testing with your devices. I think we can do that this week.

[brdr]

Awesome. Its weird that our other cisco switches that have 10Gig ports are seen correctly in Nagios after running wizard. Meaning, the WARNING and CRITICALs level show 8G and 9G respectively.

Anyway, I'm setting up a webex with your support team on another issue with Network / Router wizard (thread 33838). If you have any open questions, or want me to show anything perhaps you could join the webexe.

Thank you for your help!