Hi,
Im trying to add some WMI counters to perform a deep monitoring in SQL Servers,
I have installed an snmp extension agent (http://erwan.l.free.fr/snmptools/) in
a W2003 x64 with SQL Sever 2005 environment to test, I retrieve some WMI
counters from performance monitor by using SNMP. It works great with snmpget
and snmpwalk instructions but it a bit slowly so when I add several interfaces
for the same server it looks like the poller timeout and do not retrieve all
the data.
Is there a way to configure the timeout of the poller? Im able to walk all the
OIDs with an snmpwalk (slowly, but Im able).
Here is the output of the poller for two interfaces, one disk and one processor:
3otw-nms-01:/opt/jffnms/engine# php poller.php 12 412
09:12:29 : H 12 : Poller Start : 6 Items.
09:12:36 : H 12 : I 412 : P 1 : snmp_counter:hd_queue(.1.3..1.1): ->
buffer(): 1 (time P:6017.35 | 0.53)
09:12:42 : H 12 : I 412 : P 2 : snmp_counter:hd_time(.1.3..2.1): ->
buffer(): 2 (time P:6022.91 | 0.13)
09:12:48 : H 12 : I 412 : P 3 : snmp_counter:hd_read(.1.3..3.1): ->
buffer(): 3 (time P:6023.14 | 0.12)
09:12:54 : H 12 : I 412 : P 4 : snmp_counter:hd_write(.1.3..4.1): ->
buffer(): 4 (time P:6019.3 | 0.13)
09:12:54 : H 12 : I 412 : P 10 : no_poller(): 0 -> rrd(*): (time P:0.34
| 5.03)
09:12:54 : H 12 : I 412 : P LPD : last_poll_date(): 1293437574 ->
db(last_poll_date): 1 (time P:0.28 | 2.74)
09:12:54 : H 12 : Poller End, Total Time: 24116.95 msec.
3otw-nms-01:/opt/jffnms/engine# php poller.php 12 407
09:13:08 : H 12 : Poller Start : 4 Items.
09:13:09 : H 12 : I 407 : P 1 : snmp_counter:cpu_usage(.1.3..1.1):
0.00127998361620918 -> buffer(): 1 (time P:1073.15 | 0.59)
09:13:11 : H 12 : I 407 : P 2 : snmp_counter:cpu_queue(.1.3..2.1): 0 ->
buffer(): 2 (time P:2126.39 | 0.11)
09:13:11 : H 12 : I 407 : P 10 : no_poller(): 0 -> rrd(*):
cpu_usage:0.00127998361620918 - cpu_queue:0 (time P:0.24 | 36.33)
09:13:11 : H 12 : I 407 : P LPD : last_poll_date(): 1293437591 ->
db(last_poll_date): 1 (time P:0.58 | 2.9)
09:13:11 : H 12 : Poller End, Total Time: 3264.24 msec.
First one retrieves no data meanwhile the second one retrieves all the data,
look at the big times the poller needs to retrieve the values.
Thanks in advance,
Felipe
------------------------------------------------------------------------------
Learn how Oracle Real Application Clusters (RAC) One Node allows customers
to consolidate database storage, standardize their database environment, and,
should the need arise, upgrade to a full multi-node Oracle RAC database
without downtime or disruption
http://p.sf.net/sfu/oracle-sfdevnl
_______________________________________________
jffnms-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jffnms-users
