НР ProCurve 1810G and rddtool snmp monitoring

[Max Sorokin]

Hi All,

We have 2 HP ProCurve 1810G switches and we use rddtool to collect SNMP statistics from ports. Our problem is, that data which has been collected is incorrect and will cause a stir from the real data. With other swithces in infrastructure such as BayNetworks this tool works fine.

Is it's problem of 1810G switches or the problem is with rddtool incompability with these switches?

BR
Max

p.S. this thread has been moevd from Switches, Hubs, Modems (Legacy ITRC forum) to Web and Unmanaged - Hp forum moderator

[Mohammed Faiz]

Hi,

Could you expand a bit on what "the data is incorrect" means?
I can't comment on the 1810G's directly but I'm running cacti (which uses rrdtool) on a large range of Procurves without an issue.

[Max Sorokin]

Hi Mohhamed,

the data is incorrect means following:

we have a port on ProCurve 1810G switch, from this port we connect to port on BayNetworks switch. On BayNetworks we case see correct data on ProCurve the data i different from data on BayNetworks switch.
In attachment you can find a file "hp1810G SNMP stat .zip" there are 2 pictures: hp1810G.jpg - this is data from ProCurve switch. no_hp1810G.jpg - this is a data from BayNetwork switches. both switches measure the traffic from same source.

My english is bad and if you need more info you just ask me in simple words and i give you answers.

Max

[Mohammed Faiz]

Ok, that is very different!
Can you tell what OID you are querying for the 1810G?
Unfortunately I don't have an 1810 I can test this with for you. I would check what firmware you are running on the 1810G just in case it is a switch bug.

[Max Sorokin]

Ok,
I send to you string from configuration file which refer to the OID of the specified port:

Target[bu_1810g_24]: ifInOctets.24&ifOutOctets.24:public@192.168.1.99:::::2

Max

[Mohammed Faiz]

That should be fine. You could try changing it to ifHCInOctets/ifHCOutOctets just as a test.
What firmware version are you running on the 1810G?

[Max Sorokin]

Current firmware is: HP ProCurve 1810G - 24 GE, P.1.17, eCos-2.0

[Max Sorokin]

Good evening:

In this case SNMP statistics does not correctly collected:

Target[bu_1810g_24]: ifHCInOctets.24&ifHCOutOctets.24:public@192.168.1.99:::::2


In this case also:

Target[bu_1810g_24]: ifHCInOctets.24/ifHCOutOctets.24:public@192.168.1.99:::::2

[Mohammed Faiz]

Ok, can you try doing an snmp-get manually to see what it returns.

E.g.

> E:\net-snmp\bin>snmpget -v 2c -c public HOSTNAME 1.3.6.1.2.1.2.2.1.1
0.49

> IF-MIB::ifInOctets.49 = Counter32: 49186954

and then do the same for ifOutOctets on the baynetworks switch at the same time

> E:\net-snmp\bin>snmpget -v 2c -c public HOSTNAME 1.3.6.1.2.1.2.2.1.16.50

> IF-MIB::ifOutOctets.48 = Counter32: 49182417

[Max Sorokin]

Hi Mohammed,
answer to your last question:

root@hub:/# snmpget -v 2c -c public 192.168.1.99 1.3.6.1.2.1.2.2.1.16.50
IF-MIB::ifOutOctets.50 = No Such Instance currently exists at this OID
root@hub:/# snmpget -v 2c -c public 192.168.1.99 1.3.6.1.2.1.2.2.1.0.49
IF-MIB::ifEntry.0.49 = No Such Object available on this agent at this OID

BR,
Max

[Mohammed Faiz]

Sorry, I should have said the last number of the OID is the port on your switch. So where I had port 49:

1.3.6.1.2.1.2.2.1.10.49

You want port 24

1.3.6.1.2.1.2.2.1.10.24

And then on your Baynetworks switch:

1.3.6.1.2.1.2.2.1.16.X

where X is the port connecting to the 1810G

[Max Sorokin]

Mohammed, sorry for my poor english, but i do not understand what you want to tell me in your last post :)

[Mohammed Faiz]

No problem. Just run this:

root@hub:/# snmpget -v 2c -c public 192.168.1.99 1.3.6.1.2.1.2.2.1.16.24

Then for the BayNetworks switch:

root@hub:/# snmpget -v 2c -c public BAYNETWORKS_IP 1.3.6.1.2.1.2.2.1.16.XX

XX = port number of port connecting to HP 1810G

[Max Sorokin]

Thanks :)

[Max Sorokin]

The result is:

# snmpget -v 2c -c public 192.168.1.99 1.3.6.1.2.1.2.2.1.16.24
IF-MIB::ifOutOctets.24 = Counter32: 2121687475
root@hub:~# snmpget -v 2c -c public 192.168.1.72 1.3.6.1.2.1.2.2.1.16.13
Timeout: No Response from 192.168.1.72.
root@hub:~# snmpget -v1 -c public 192.168.1.72 1.3.6.1.2.1.2.2.1.16.13
IF-MIB::ifOutOctets.13 = Counter32: 452595655

[Mohammed Faiz]

Ok, that's working, can you do:

# snmpget -v 2c -c public 192.168.1.99 1.3.6.1.2.1.2.2.1.10.24

# snmpget -v 2c -c public 192.168.1.72 1.3.6.1.2.1.2.2.1.16.13

then repeat :

# snmpget -v 2c -c public 192.168.1.99 1.3.6.1.2.1.2.2.1.10.24

# snmpget -v 2c -c public 192.168.1.72 1.3.6.1.2.1.2.2.1.16.13

Then we can compare the values directly.

[Max Sorokin]

The result is:

# snmpget -v 2c -c public 192.168.1.99 1.3.6.1.2.1.2.2.1.16.24
IF-MIB::ifOutOctets.24 = Counter32: 1703719157
root@hub:~# snmpget -v1 -c public 192.168.1.72 1.3.6.1.2.1.2.2.1.16.13
IF-MIB::ifOutOctets.13 = Counter32: 1435926472
root@hub:~# snmpget -v 2c -c public 192.168.1.99 1.3.6.1.2.1.2.2.1.16.24
IF-MIB::ifOutOctets.24 = Counter32: 1740820409
root@hub:~# snmpget -v1 -c public 192.168.1.72 1.3.6.1.2.1.2.2.1.16.13
IF-MIB::ifOutOctets.13 = Counter32: 1443762294

[Mohammed Faiz]

So,

1740820409 - 1703719157 = 37101252 bytes

1443762294 - 1435926472 = 7835822 bytes

It's definitely not a problem with rrdtool.

I'm not sure what would cause a mis-match between the values like that.
Have you checked the ports for any errors or a speed/duplex mismatch?

[Max Sorokin]

Thanks a lot :)
Let's ask in other manner,
Why only from HP1810G we have incorrect information, while from
HP-1700-8
HP-1700-24
HP-2610
Bay-350
Bay-450
Cisco 3750
AT-750FS-24
3Com1100
3Com4400
PCrouter Linux
Mikrotik RB750,RB1100
all data displayed correctly and collection method we use the same.
In the attachment i sent to you statistics from different ports of HP1810G.

All ports works correctly, there are no errors on ports, no problems with speed.
Only one problem: port statistics did not diplayed correctly, therefore we can not
count the ammount of the consumed traffic.

BR,
Max

[Mohammed Faiz]

I'm afraid there's not much else I can suggest.
Open a support call with HP support but before you do that you should:

- Do a factory reset of the switch
- Re-apply the firmware image

hth

[Brad Morton]

I had the same problem. Talked to HP..they were no help. Basicly said it's not us it's you.

[AlexanderSwen]

hi Mohammed,

I seem to have a similar problem. (using munin snmp plugin)

Procurve 1810g - 24 GE, P.2.2, eCos-2.0, CFE-2.1

Image1 P.1.20   Image2 P.2.2 default image

(so I guess I've got 2.2 as the firmware image)

I've got a problem and i suspect the snmp daemon on the switch to be not as complete as it should be. this is what my snmp plugin tells me:

 " NOTE! This switch supports only 32 bit byte counters which makes the plugin unreliable and unsuitable for most 100Mb (or faster) interfaces, where bursts are expected to exceed 50Mbps.  This means that for interfaces where much traffic is sent this plugin will report false thruputs and cannot be trusted."

# snmpwalk -v2c -c public switch ifHCInOctets.12 -O fn
.1.3.6.1.2.1.31.1.1.1.6.12 = No Such Instance currently exists at this OID
# snmpwalk -v2c -c public switch ifInOctets.12 -O fn
.1.3.6.1.2.1.2.2.1.10.12 = Counter32: 28202

 If you have got any influence on the HP switch support team, please ask them to increase the info we can query from this switch.

regards,

Alex