nv-l
[Top] [All Lists]

RE: [nv-l] HSRP interface gets deleted and recreated every minute

To: <nv-l@lists.us.ibm.com>
Subject: RE: [nv-l] HSRP interface gets deleted and recreated every minute
From: "Barr, Scott" <Scott_Barr@csgsystems.com>
Date: Thu, 18 Mar 2004 08:33:48 -0600
Delivery-date: Thu, 18 Mar 2004 14:42:27 +0000
Envelope-to: nv-l-archive@lists.skills-1st.co.uk
Reply-to: nv-l@lists.us.ibm.com
Sender: owner-nv-l@lists.us.ibm.com
Thread-index: AcQMTXPgNUiXr+WsQIWvWLMxugUQ4QAqIsvw
Thread-topic: [nv-l] HSRP interface gets deleted and recreated every minute
Make sure -S is on your netmon in /usr/OV/conf/ovsuf

-----Original Message-----
From: owner-nv-l@lists.us.ibm.com [mailto:owner-nv-l@lists.us.ibm.com]On
Behalf Of Francois Le Hir
Sent: Wednesday, March 17, 2004 12:25 PM
To: nv-l@lists.us.ibm.com
Subject: [nv-l] HSRP interface gets deleted and recreated every minute


Netview 7.1.4 and TEC 3.9 on AIX 5.1

I integrated Netview with TEC and found out that I was receiving a high
number of interface down for a few interfaces (always the same and about
every minute). After searching the logs on netview I saw that theses
interfaces are HSRP interfaces. Netview detect a HSRP switch over of the
interface, send an interface down interface and proceed with the deletion
of the interface.

Q1: why is netview sending an interface down event when it detect an HSRP
switch over ?
Q2: why is the deletion of the interface in netview not seen by TEC as a
closing event ? (I use the standard TEC_ITS.rs and netview.rls) Probably
this should be part of the logic in netview.rls on the TEC side.

After searching further in the netview log, I saw in netmon.trace that the
interface is always recreated in the same switch:

Netview see the HSRP switch over, send a down event and delete the
interface

15:09:26 : .//nl_snmper.c[319] : sending SNMP to 10.4.6.2 op = HSRP req =
SysName reqid = 1334547
15:09:26 : .//nl_snmper.c[1146] : recv_snmp: from
aclcyul02ci412.aircanada.ca (10.4.6.2) op = HSRP req = SysName reqid =
1334547
15:09:26 : .//nl_snmpstate.c[3651] : actonHsrpSysName: Node
aclcyul02ci412.aircanada.ca snmp addr 10.4.6.2 old sysName
(ACLCYUL02CI412.aircanada.ca) new sysName (ACLCYUL02CI411.aircanada.ca)
15:09:26 : .//nl_snmpstate.c[3686] : HSRP Switch: Node
aclcyul02ci412.aircanada.ca snmp addr 10.4.6.2 old sysName
(ACLCYUL02CI412.aircanada.ca) new sysName (ACLCYUL02CI411.aircanada.ca)
15:09:26 : .//nl_event.c[333] : DOWN event: 10.4.6.2
(aclcyul02ci412.aircanada.ca)
15:09:26 : .//nl_event.c[632] : HSRP interface 10.4.6.2 is deleted from
aclcyul02ci412.aircanada.ca
15:09:26 : .//nl_fixup.c[137] : fixupIfaceSnmpConf() for 10.4.6.2
15:09:26 : .//nl_snmpstate.c[2752] : endSnmpOperation HSRP_POLL: Node
aclcyul02ci412.aircanada.ca: polling next HSRP interface 10.4.11.2


Netview rediscover the HSRP interface:

15:09:38 : .//nl_pinger.c[375] : sending ping to 10.4.6.2 seqnum = 41472
ident = 57482 timeout = 2
15:09:39 : .//nl_pinger.c[1505] : -> received ping, ident=57482 seq=41472,
from 10.4.6.2 ()
15:09:39 : .//nl_pinger.c[1079] : calling allocIfObjectId() :
ifaddr=10.4.6.2
15:09:39 : DEBUG: Get Unique Name For: Interface:10.4.6.2
15:09:39 : DEBUG: Create Object: 10.4.6.2
15:09:40 : .//nl_pinger.c[1992] : addNode: 10.4.6.2
(aclcyul02ci412.aircanada.ca)
15:09:40 : .//nl_pinger.c[2156] : Merging interface 10.4.6.2 with node
aclcyul02ci412.aircanada.ca
15:10:11 : .//nl_fixup.c[476] : fixupNmNodeSnmpConfEx() for
aclcyul02ci412.aircanada.ca
15:10:12 : .//nl_fixup.c[476] : fixupNmNodeSnmpConfEx() for
aclcyul02ci412.aircanada.ca
15:10:14 : .//nl_snmpstate.c[4741] : ### atNetAddress = 10.4.6.2,
atPhysAddress = 0x00000C07AC02, atIfIndex = 3
15:10:14 : .//nl_snmpstate.c[4810] : actonSecondaryIf: Found HSRP 10.4.6.2,
best guess so far on aclcyul02ci412.aircanada.ca, physaddr 0x00000C07AC02
15:10:15 : .//nl_event.c[144] : changeIfEvent:  node 10.4.6.2 eventnum ip
Mask Change
Node 'aclcyul02ci412.aircanada.ca' is in correct network
15:10:15 : .//nl_event.c[144] : changeIfEvent:  node 10.4.6.2 eventnum if
type Change
15:10:15 : .//nl_event.c[144] : changeIfEvent:  node 10.4.6.2 eventnum if
descr Change
15:10:15 : .//nl_snmpstate.c[7301] : endIfTable: Node
aclcyul02ci412.aircanada.ca: updating HSRP interface 10.4.6.2
15:10:15 : .//nl_event.c[611] : HSRP interface 10.4.6.2 is added to node
aclcyul02ci412.aircanada.ca
15:10:15 : .//nl_event.c[144] : changeIfEvent:  node 10.4.6.2 eventnum if
Alias Change

I see the lines

15:09:39 : DEBUG: Get Unique Name For: Interface:10.4.6.2
15:09:39 : DEBUG: Create Object: 10.4.6.2
15:09:40 : .//nl_pinger.c[1992] : addNode: 10.4.6.2
(aclcyul02ci412.aircanada.ca)

Q3: does it means that because the name resolution for 10.4.6.2 is
aclcyul02ci412.aircanada.ca, then the interface is going to be recreated in
the same switch as the one from where it was just deleted ? It should be
recreated in the other switch (aclcyul02ci411)

Q4: What is supposed to be the name resolution for an HSRP interface ? By
default presently I resolve it to the name of the switch that is primary
for HSRP. If I were to resolve it to an uniq name (deferent than the name
of the primary or secondary HSRP switch), I would get device in the netview
database that would get discovered by netview with the name of the HSRP
interface (ie most often the first interface seen by netview when
discovering a device) and this would be even worst a problem because a lot
of my correlation logic rely on the selection name of a device.

Thanks,
Salutations, / Regards,

Francois Le Hir
Network Projects & Consulting Services
IBM Global Services
Phone: (514) 205 6695



<Prev in Thread] Current Thread [Next in Thread>

Archive operated by Skills 1st Ltd

See also: The NetView Web