windows xp disk usage - down peaks !!

[intenso]

hi,

i start to graph five new servers with cacti. the network and cpu graphs shows very fine but the disk usage graphs have same down peaks - i dont know why!!

my cacti-system:
debian-sarge, cacti 0.8.6.h, cactid 0.8.6.e

the "graphd"-system:
windows xp sp2, snmp and wmi support on

have any one an idea?

[BSOD2600]

Looks valid to me. Seems between 14:00 to 17:00 you were downloading or something, since you've got spikes in bandwidth, cpu, disk, and memory. You're saying your disk space on that drive doesnt change by that much?

[intenso]

the variable values (used RAM, etc) can change, but why change the fix falues (total physical RAM, total virtual RAM, etc.)?

here are two possibilities to fix this problem that i found:

a)
in a onther thread i have seen, that the snmp timeout can be increased to fix this problem.
cacti.log:
SNMP Timeout detected [500] on Host...

but if i change the snmp timeout value in cacti to 2000 or 10000 or whatever, the same message is in the cacti log.

b)
i have a old cactid (086e). i have a debian system but when i compile the cactid i can not run make!?
i run ./configure and then i would run make.
the ./configure runs correctly to the end.

sorry for my stupid english and my descriptions

[TheWitness]

You can reduce the number of OID's grabbed by Cacti in one poll by going to Settings->Poller->Max OID's Per Request. Make it something smaller. This might help. Also, you may benefit from the latest version of Cactid. I have a SVN binary floating around somewhere from just a few days ago. If you look for something like "cactid.zip" on the forum and look for a date in the last few weeks, you might have a good copy.

TheWitness