[Check_mk (english)] Having *inventory* use snmpwalk, but conduct actual monitoring with bulkwalks

Hi,

  we have noticed that when check_mk inventory is run, or when doing discovery from WATO, the discovery eventually times out.

  However, if we get the discovery to go through, monitoring actually works fine.

  Anyway, it seems that the problem is caused by SNMP bulkwalks against certain systems (such as Cisco CUCM). They don't seem to like "open-ended" bulkwalks where you probe every OID in existence. However, setting the host to legacy snmpv1 mode allows discovery to be conducted.

  However, during actual monitoring (when you know which OIDs you are polling, and not every check in existence), bulkwalks work just fine.

  Is there any way to have CMK not use bulkwalks for discovery, but use them during monitoring?

  Thanks.

···

--
- Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
- Vintor Oy, Itsehallintokuja 6, 02600 Espoo | www.vintor.fi -

Use the 'nobulk' tag to force Check_MK to use 'snmpwalk' instead of 'snmpbulkwalk' on the hosts you assign it to.

https://mathias-kettner.de/checkmk_snmp.html

···

-----Original Message-----
From: checkmk-en-bounces@lists.mathias-kettner.de [mailto:checkmk-en-bounces@lists.mathias-kettner.de] On Behalf Of Mäkelä, Antti
Sent: Tuesday, June 21, 2016 6:12 AM
To: checkmk-en@lists.mathias-kettner.de
Subject: [Check_mk (english)] Having *inventory* use snmpwalk, but conduct actual monitoring with bulkwalks

Hi,

  we have noticed that when check_mk inventory is run, or when doing discovery from WATO, the discovery eventually times out.

  However, if we get the discovery to go through, monitoring actually works fine.

  Anyway, it seems that the problem is caused by SNMP bulkwalks against certain systems (such as Cisco CUCM). They don't seem to like "open-ended" bulkwalks where you probe every OID in existence. However, setting the host to legacy snmpv1 mode allows discovery to be conducted.

  However, during actual monitoring (when you know which OIDs you are polling, and not every check in existence), bulkwalks work just fine.

  Is there any way to have CMK not use bulkwalks for discovery, but use them during monitoring?

  Thanks.

--
- Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
- Vintor Oy, Itsehallintokuja 6, 02600 Espoo | www.vintor.fi -

_______________________________________________
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en

Problem is that to my understanding this approach also changes the monitoring to use "nobulk". I just need the discovery/Check_MK Inventory service to use the "nobulk"...

···

--
- Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
- Vintor Oy, Itsehallintokuja 6, 02600 Espoo | www.vintor.fi -

________________________________________
From: Stier, Matthew <Matthew.Stier@us.fujitsu.com>
Sent: Tuesday, June 21, 2016 18:07
To: Mäkelä, Antti; checkmk-en@lists.mathias-kettner.de
Subject: RE: Having *inventory* use snmpwalk, but conduct actual monitoring with bulkwalks

Use the 'nobulk' tag to force Check_MK to use 'snmpwalk' instead of 'snmpbulkwalk' on the hosts you assign it to.

https://mathias-kettner.de/checkmk_snmp.html

-----Original Message-----
From: checkmk-en-bounces@lists.mathias-kettner.de [mailto:checkmk-en-bounces@lists.mathias-kettner.de] On Behalf Of Mäkelä, Antti
Sent: Tuesday, June 21, 2016 6:12 AM
To: checkmk-en@lists.mathias-kettner.de
Subject: [Check_mk (english)] Having *inventory* use snmpwalk, but conduct actual monitoring with bulkwalks

Hi,

  we have noticed that when check_mk inventory is run, or when doing discovery from WATO, the discovery eventually times out.

  However, if we get the discovery to go through, monitoring actually works fine.

  Anyway, it seems that the problem is caused by SNMP bulkwalks against certain systems (such as Cisco CUCM). They don't seem to like "open-ended" bulkwalks where you probe every OID in existence. However, setting the host to legacy snmpv1 mode allows discovery to be conducted.

  However, during actual monitoring (when you know which OIDs you are polling, and not every check in existence), bulkwalks work just fine.

  Is there any way to have CMK not use bulkwalks for discovery, but use them during monitoring?

  Thanks.

--
- Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
- Vintor Oy, Itsehallintokuja 6, 02600 Espoo | www.vintor.fi -

_______________________________________________
checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en

Hi Antti,

it’s impossible to use different features for discovery and checks. You have to use bulk or nobulk for both of them.

If bulk raised issues I’d use nobulk as fixed configuration for affected hosts.

Regards,
Marcel

···

Mäkelä, Antti Antti.Makela@vintor.fi schrieb am Di., 21. Juni 2016 17:09:

Problem is that to my understanding this approach also changes the monitoring to use “nobulk”. I just need the discovery/Check_MK Inventory service to use the “nobulk”…

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -

  • Vintor Oy, Itsehallintokuja 6, 02600 Espoo | www.vintor.fi -


From: Stier, Matthew Matthew.Stier@us.fujitsu.com

Sent: Tuesday, June 21, 2016 18:07

To: Mäkelä, Antti; checkmk-en@lists.mathias-kettner.de

Subject: RE: Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Use the ‘nobulk’ tag to force Check_MK to use ‘snmpwalk’ instead of ‘snmpbulkwalk’ on the hosts you assign it to.

https://mathias-kettner.de/checkmk_snmp.html

-----Original Message-----

From: checkmk-en-bounces@lists.mathias-kettner.de [mailto:checkmk-en-bounces@lists.mathias-kettner.de] On Behalf Of Mäkelä, Antti

Sent: Tuesday, June 21, 2016 6:12 AM

To: checkmk-en@lists.mathias-kettner.de

Subject: [Check_mk (english)] Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Hi,

we have noticed that when check_mk inventory is run, or when doing discovery from WATO, the discovery eventually times out.

However, if we get the discovery to go through, monitoring actually works fine.

Anyway, it seems that the problem is caused by SNMP bulkwalks against certain systems (such as Cisco CUCM). They don’t seem to like “open-ended” bulkwalks where you probe every OID in existence. However, setting the host to legacy snmpv1 mode allows discovery to be conducted.

However, during actual monitoring (when you know which OIDs you are polling, and not every check in existence), bulkwalks work just fine.

Is there any way to have CMK not use bulkwalks for discovery, but use them during monitoring?

Thanks.

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -

  • Vintor Oy, Itsehallintokuja 6, 02600 Espoo | www.vintor.fi -


checkmk-en mailing list

checkmk-en@lists.mathias-kettner.de

http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en


checkmk-en mailing list

checkmk-en@lists.mathias-kettner.de

http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en

Ok, well, if that’s not possible then we’ll just have to live with it.

I’ve found a related issue where it seems that I cannot turn off bulkwalk for SNMPv3. Setting “nobulk” works if using SNMPv1/SNMPv2c, but looks
like SNMPv3 is stuck with bulkwalk (at least according to cmk –vvvvvvv –debug –I). Oh well, for most part the enhanced security of SNMPv3 is not strictly necessary in our environment.

···

Mäkelä, Antti Antti.Makela@vintor.fi schrieb am Di., 21. Juni 2016 17:09:

Problem is that to my understanding this approach also changes the monitoring to use “nobulk”. I just need the discovery/Check_MK Inventory service to use the “nobulk”…

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
  • Vintor Oy, Itsehallintokuja 6, 02600 Espoo |
    www.vintor.fi
    -

From: Stier, Matthew Matthew.Stier@us.fujitsu.com
Sent: Tuesday, June 21, 2016 18:07
To: Mäkelä, Antti;
checkmk-en@lists.mathias-kettner.de
Subject: RE: Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Use the ‘nobulk’ tag to force Check_MK to use ‘snmpwalk’ instead of ‘snmpbulkwalk’ on the hosts you assign it to.

https://mathias-kettner.de/checkmk_snmp.html

-----Original Message-----
From:
checkmk-en-bounces@lists.mathias-kettner.de [mailto:checkmk-en-bounces@lists.mathias-kettner.de] On Behalf Of Mäkelä, Antti
Sent: Tuesday, June 21, 2016 6:12 AM
To: checkmk-en@lists.mathias-kettner.de
Subject: [Check_mk (english)] Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Hi,

we have noticed that when check_mk inventory is run, or when doing discovery from WATO, the discovery eventually times out.

However, if we get the discovery to go through, monitoring actually works fine.

Anyway, it seems that the problem is caused by SNMP bulkwalks against certain systems (such as Cisco CUCM). They don’t seem to like “open-ended” bulkwalks where you probe every OID in existence. However, setting the host to legacy snmpv1 mode allows discovery
to be conducted.

However, during actual monitoring (when you know which OIDs you are polling, and not every check in existence), bulkwalks work just fine.

Is there any way to have CMK not use bulkwalks for discovery, but use them during monitoring?

Thanks.

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
  • Vintor Oy, Itsehallintokuja 6, 02600 Espoo |
    www.vintor.fi
    -

checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en


checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en

Hi Antti,

which version are you using? Regarding to Werk#2494 (https://mathias-kettner.de/check_mk_werks.php?werk_id=2494&HTML=yes) the bulkwalk issue with SNMPv3 should be fixed since 1.2.6p10 and in current stable (since 1.2.7i3)…

Regards,

Marcel

···

Mäkelä, Antti Antti.Makela@vintor.fi schrieb am Mi., 22. Juni 2016 um 08:57 Uhr:

Ok, well, if that’s not possible then we’ll just have to live with it.

I’ve found a related issue where it seems that I cannot turn off bulkwalk for SNMPv3. Setting “nobulk” works if using SNMPv1/SNMPv2c, but looks
like SNMPv3 is stuck with bulkwalk (at least according to cmk –vvvvvvv –debug –I). Oh well, for most part the enhanced security of SNMPv3 is not strictly necessary in our environment.

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -

From: Marcel Schulte [mailto:schulte.marcel@gmail.com]
Sent: 21. kesäkuuta 2016 18:17
To: Mäkelä, Antti; Stier, Matthew; checkmk-en@lists.mathias-kettner.de
Subject: Re: [Check_mk (english)] Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Hi Antti,

it’s impossible to use different features for discovery and checks. You have to use bulk or nobulk for both of them.

If bulk raised issues I’d use nobulk as fixed configuration for affected hosts.

Regards,
Marcel

Mäkelä, Antti Antti.Makela@vintor.fi schrieb am Di., 21. Juni 2016 17:09:

Problem is that to my understanding this approach also changes the monitoring to use “nobulk”. I just need the discovery/Check_MK Inventory service to use the “nobulk”…

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
  • Vintor Oy, Itsehallintokuja 6, 02600 Espoo |
    www.vintor.fi
    -

From: Stier, Matthew Matthew.Stier@us.fujitsu.com
Sent: Tuesday, June 21, 2016 18:07
To: Mäkelä, Antti;
checkmk-en@lists.mathias-kettner.de
Subject: RE: Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Use the ‘nobulk’ tag to force Check_MK to use ‘snmpwalk’ instead of ‘snmpbulkwalk’ on the hosts you assign it to.

https://mathias-kettner.de/checkmk_snmp.html

-----Original Message-----
From:
checkmk-en-bounces@lists.mathias-kettner.de [mailto:checkmk-en-bounces@lists.mathias-kettner.de] On Behalf Of Mäkelä, Antti
Sent: Tuesday, June 21, 2016 6:12 AM
To: checkmk-en@lists.mathias-kettner.de
Subject: [Check_mk (english)] Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Hi,

we have noticed that when check_mk inventory is run, or when doing discovery from WATO, the discovery eventually times out.

However, if we get the discovery to go through, monitoring actually works fine.

Anyway, it seems that the problem is caused by SNMP bulkwalks against certain systems (such as Cisco CUCM). They don’t seem to like “open-ended” bulkwalks where you probe every OID in existence. However, setting the host to legacy snmpv1 mode allows discovery
to be conducted.

However, during actual monitoring (when you know which OIDs you are polling, and not every check in existence), bulkwalks work just fine.

Is there any way to have CMK not use bulkwalks for discovery, but use them during monitoring?

Thanks.

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
  • Vintor Oy, Itsehallintokuja 6, 02600 Espoo |
    www.vintor.fi
    -

checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en


checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en

This is 1.2.8p4…

Granted, I didn’t use tags directly in config file, but I set the ”don’t use bulk” via WATO. Basically, I set “Hosts using SNMP bulk walk (enforces
SNMP v2c)” to NOT include my target hosts. Is there a separate place to disable bulkwalk for SNMPv3?

···

Hi Antti,

which version are you using? Regarding to Werk#2494 (https://mathias-kettner.de/check_mk_werks.php?werk_id=2494&HTML=yes ) the bulkwalk issue with SNMPv3 should
be fixed since 1.2.6p10 and in current stable (since 1.2.7i3)…

Regards,

Marcel

Mäkelä, Antti Antti.Makela@vintor.fi schrieb am Mi., 22. Juni 2016 um 08:57 Uhr:

Ok, well, if that’s not possible then we’ll just have to live with it.

I’ve found a related issue where it seems that I cannot turn off bulkwalk for SNMPv3.
Setting “nobulk” works if using SNMPv1/SNMPv2c, but looks like SNMPv3 is stuck with bulkwalk (at least according to cmk –vvvvvvv –debug –I). Oh well, for most part the enhanced security of SNMPv3 is not strictly necessary in our environment.

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -

From: Marcel
Schulte [mailto:schulte.marcel@gmail.com]
Sent: 21. kesäkuuta 2016 18:17
To: Mäkelä, Antti; Stier, Matthew;
checkmk-en@lists.mathias-kettner.de
Subject: Re: [Check_mk (english)] Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Hi Antti,

it’s impossible to use different features for discovery and checks. You have to use bulk or nobulk for both of them.

If bulk raised issues I’d use nobulk as fixed configuration for affected hosts.

Regards,
Marcel

Mäkelä, Antti Antti.Makela@vintor.fi schrieb am Di., 21. Juni 2016 17:09:

Problem is that to my understanding this approach also changes the monitoring to use “nobulk”. I just need the discovery/Check_MK Inventory service to use the “nobulk”…

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
  • Vintor Oy, Itsehallintokuja 6, 02600 Espoo |
    www.vintor.fi
    -

From: Stier, Matthew Matthew.Stier@us.fujitsu.com
Sent: Tuesday, June 21, 2016 18:07
To: Mäkelä, Antti;
checkmk-en@lists.mathias-kettner.de
Subject: RE: Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Use the ‘nobulk’ tag to force Check_MK to use ‘snmpwalk’ instead of ‘snmpbulkwalk’ on the hosts you assign it to.

https://mathias-kettner.de/checkmk_snmp.html

-----Original Message-----
From:
checkmk-en-bounces@lists.mathias-kettner.de [mailto:checkmk-en-bounces@lists.mathias-kettner.de] On Behalf Of Mäkelä, Antti
Sent: Tuesday, June 21, 2016 6:12 AM
To: checkmk-en@lists.mathias-kettner.de
Subject: [Check_mk (english)] Having inventory use snmpwalk, but conduct actual monitoring with bulkwalks

Hi,

we have noticed that when check_mk inventory is run, or when doing discovery from WATO, the discovery eventually times out.

However, if we get the discovery to go through, monitoring actually works fine.

Anyway, it seems that the problem is caused by SNMP bulkwalks against certain systems (such as Cisco CUCM). They don’t seem to like “open-ended” bulkwalks where you probe every OID in existence. However, setting the host to legacy snmpv1 mode allows discovery
to be conducted.

However, during actual monitoring (when you know which OIDs you are polling, and not every check in existence), bulkwalks work just fine.

Is there any way to have CMK not use bulkwalks for discovery, but use them during monitoring?

Thanks.

  • Dr. Antti Mäkelä | Senior Architect | CCIE #20962 -
  • Vintor Oy, Itsehallintokuja 6, 02600 Espoo |
    www.vintor.fi
    -

checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en


checkmk-en mailing list
checkmk-en@lists.mathias-kettner.de
http://lists.mathias-kettner.de/mailman/listinfo/checkmk-en