|
![]() |
|
| Author |
|
|||||||
|
natslovR
Posts: 1686
Location: Canberra, Australian Capital Territory
|
I want to set windows performance counter logs to roll over every day and start a new log when they do.
the problem is windows performance logs can just stop when something goes wrong and need to be manually restarted. does anyone know a non manual way to tell when a perfcounter log is not working I use to roll them hourly, and scan eventlog for hourly restart events, but I don't want them to be down for more than a day which is what that will reqire once they are set to daily rollover. |
|||||||
| #0 03:23pm 27/06/08 |
|
|||||||
|
system
|
--
|
|||||||
| #0 |
|
|||||||
|
TicMan
Posts: 3420
Location: Brisbane, Queensland
|
Have no idea, but have you thought about using SNMP monitoring (Cacti) instead of performance counter?
|
|||||||
| #1 02:05pm 27/06/08 |
|
|||||||
|
natslovR
Posts: 1687
Location: Canberra, Australian Capital Territory
|
I haven't found anything that beats the simplicity and flexibility of raw perflogs.
even nt4 perflogs are still supported by relog and logparser, so it seems to me to be the most future proof format. |
|||||||
| #2 03:09pm 27/06/08 |
|
|||||||
|
TicMan
Posts: 3421
Location: Brisbane, Queensland
|
What about writing your own perf counters -> MySQL/MSSQL script which retrieves the values via WMI and then spits it into the database? You can then run reports or graphing software against the values stored within the database.
|
|||||||
| #3 03:23pm 27/06/08 |
|
|||||||
|
natslovR
Posts: 5779
Location: Sydney, New South Wales
|
i'm trying to keep the performance data in a format that's no hassle to archive and withstands the test of time.
raw perflogs can be easily loaded in to a db to analyse and pretty pics can be churned out and useful warnings given using tools like PAL http://codeplex.com/PAL anything i write myself is only going to last as long as i'm there to maintain it, and an archiving process will need to be developed to put the data back to the filesystem anyway. |
|||||||
| #4 10:34pm 27/06/08 |
|
|||||||
|
system
|
--
|
|||||||
| #4 |
|
|||||||
|
| ||||||||