event 0x10d80 Problem detected with data in the Interface Performance Calculations

book

Article ID: 141893

calendar_today

Updated On:

Products

CA eHealth CA Spectrum

Issue/Introduction

For some interfaces of specific devices, you may get often the the event: 0x10d80 that may affect the  IN_LOAD and OUT_LOAD monitoring.

                Problem detected with data in the Interface Performance Calculations

SOURCE DATA ************************************

Initial X_IN_OCTETS: 3690255947302006
Final X_IN_OCTETS: 3690261998657745

Initial X_OUT_OCTETS: 3936512669608771
Final X_OUT_OCTETS: 3936532747070030

Initial X_IN_ERROR_PKTS: 2
Final X_IN_ERROR_PKTS: 2

Initial X_OUT_ERROR_PKTS: 0
Final X_OUT_ERROR_PKTS: 0

Initial X_IN_NO_RESOURCE_PKTS: 0
Final X_IN_NO_RESOURCE_PKTS: 0

Initial X_OUT_NO_RESOURCE_PKTS: 0
Final X_OUT_NO_RESOURCE_PKTS: 0

Initial X_IN_NUCAST_PKTS: 41193845
Final X_IN_NUCAST_PKTS: 41193872

Initial X_OUT_NUCAST_PKTS: 3084851016
Final X_OUT_NUCAST_PKTS: 3084851044

Initial X_IN_MCAST_PKTS: 41193828
Final X_IN_MCAST_PKTS: 41193855

Initial X_OUT_MCAST_PKTS: 3084851004
Final X_OUT_MCAST_PKTS: 3084851032

Initial X_IN_BCAST_PKTS: 17
Final X_IN_BCAST_PKTS: 17

Initial X_OUT_BCAST_PKTS: 12
Final X_OUT_BCAST_PKTS: 12

Initial X_IN_UCAST_PKTS: 3409694604525
Final X_IN_UCAST_PKTS: 3409701142611

Initial X_OUT_UCAST_PKTS: 3745949816989
Final X_OUT_UCAST_PKTS: 3745965376977

Initial X_IN_UNKNOWN_PROTOS: 0
Final X_IN_UNKNOWN_PROTOS: 0

Initial X_SYSTEM_UP_TIME: 3542613419
Final X_SYSTEM_UP_TIME: 3542614427
Delta SYSTEM_UP_TIME: 1008

PREVIOUS IFTYPE: 6
CURRENT IFTYPE: 6

PREVIOUS X_LOW_BANDWIDTH: 4294967295
CURRENT X_LOW_BANDWIDTH: 4294967295

PREVIOUS X_HIGH_BANDWIDTH: 10000
CURRENT X_HIGH_BANDWIDTH: 10000


CALCULATED METRICS ***************************************

PREVIOUS BIT_RATE_IN: 0
PREVIOUS BIT_RATE_OUT: 0

CURRENT BIT_RATE_IN: 4802663284
CURRENT BIT_RATE_OUT: 15934493062


PREVIOUS LOAD_TOTAL: 0
PREVIOUS LOAD_IN: 0
PREVIOUS LOAD_OUT: 0

CURRENT LOAD_TOTAL: 207
CURRENT LOAD_IN: 48
CURRENT LOAD_OUT: 159


PREVIOUS IN_THROUGHPUT_PKTS: 0
PREVIOUS OUT_THROUGHPUT_PKTS: 0
PREVIOUS TOTAL_THROUGHPUT_PKTS: 0

CURRENT IN_THROUGHPUT_PKTS: 648622
CURRENT OUT_THROUGHPUT_PKTS: 1543652
CURRENT TOTAL_THROUGHPUT_PKTS: 2192274


PREVIOUS PERCENT_ERROR_PKTS: 0
CURRENT PERCENT_ERROR_PKTS: 0

PREVIOUS PERCENT_IN_ERRORS: 0
CURRENT PERCENT_IN_ERRORS: 0

PREVIOUS PERCENT_OUT_ERRORS: 0
CURRENT PERCENT_OUT_ERRORS: 0


PREVIOUS PERCENT_DISCARDED_PKTS: 0
CURRENT PERCENT_DISCARDED_PKTS: 0

PREVIOUS PERCENT_IN_DISCARDS: 0
CURRENT PERCENT_IN_DISCARDS: 0

PREVIOUS PERCENT_OUT_DISCARDS: 0
CURRENT PERCENT_OUT_DISCARDS: 0


INTERNAL SOURCES ATTRIBUTES ******************************

0x12daa
0x12dab
0x12dac
0x12db6
0x12db2
0x12db4
0x12db8
0x12dad
0x12db7
0x12db3
0x12db5
0x130d1
0x12dfa
0x12db0
0x12dae
0x12db1
0x12daf
0x12bb5
0x12bb4
0x220177
0x12df9

EXTERNAL SOURCES ATTRIBUTES:

0x11f65
0x11f69
0x11f66
0x11317
0x11318
0x11319
0x1131a
0x11f6a
0x1131e
0x1131f
0x11321
0x11348
0x1134c
0x11f6c
0x11f6d
0x11f70
0x11f71
0x11f6b
0x22001a
0x22001a
0x22005e 

Which are the possible causes of that message? Any how to prevent it from occurring?

    

Environment

Release : 10.3

Component : Spectrum Core / SpectroSERVER

Resolution

One of the possible cause of 0x10d80 is that some X_IN value is > than the corresponding X_OUT value (see Article Id: 130231), but if also the data is consistent, per Spectrum design, the SpectroSERVER raises 0x10d80 event when current LOAD_TOTAL(LOAD_IN + LOAD_OUT) value exceeds default threshold of 200.

Playing with the .vnmrc parameter port_perf_valid_result_age may help to reduce the frequency of 0x10d80

During port performance calculation two attributes read from .vnmrc file. The first attribute is  port_perf_valid_result_age (Default 8 sec) and the second one is port_perf_node_ageout (Default 900 secs).
It you configure the port_perf_valid_result_age attribute value to 60 secs this means spectrum polls the interface only once every 60 secs.
If any request comes from oneClick during valid time age, spectrum simply returns already calculated metrics (cached results) instead of new polling on the interface. 

The other attribute port_perf_node_ageout value is used for identifying the stale node of an interface in case polling fails.