[solved]"Host MIB - Available Disk Space" shows wrong values

[mgtech]

Hello,

a few info about my cacti installation:
- Version 0.8.7g running on Linux
- NET-SNMP version: 5.4.2.1
- RRDTool 1.3.x

My Cacti-Box queries mostly Windows-based servers, some are Windows 2003, others are Windows 2008 (both 1st edition and R2). I normally use the following data queries:
1) SNMP - Get Mounted Partitions
2) SNMP - Get Processor Information
3) SNMP - Interface Statistics

Now let's talk about the problem... data queries 2 and 3 are working ok.
About data query 1, many of the Windows 2003 servers (but not all of them) show the correct graph, while almost all of the 2008 servers are showing wrong graphs.

I use that data query in order to monitor disk space of the Hard Disk, RAM and Virtual Memory (they are seen as disks by SNMP). The problem is that very ofter the query return the wrong value.

Just take a look to the attached graphs: you can see that sometimes the values are correct, and sometimes they are not. Of course the Physical Memory size won't change from 6.0GB to 0 and then to 6GB again many times per day. And still it's obviously not possible that used memory (red colour) is greather than total memory (blue colour).

The same situation is for Virtual Memory, and the disks. Please note that the drive letters and labels are correct (I just masked them on the images).

Anybody can be of help?

Thanks in advance.

[mgtech]

A few more info: in the log I can see a lots of timeout...
06/13/2011 08:45:13 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'13', and OID:'.1.3.6.1.2.1.25.2.3.1.6.6'
06/13/2011 08:45:13 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'12', and OID:'.1.3.6.1.2.1.25.2.3.1.5.6'
06/13/2011 08:46:09 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'14', and OID:'.1.3.6.1.2.1.25.2.3.1.5.2'
06/13/2011 08:47:05 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'03', and OID:'.1.3.6.1.2.1.25.2.3.1.6.2'
06/13/2011 08:47:06 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'06', and OID:'.1.3.6.1.2.1.25.2.3.1.6.2'
06/13/2011 08:47:06 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'05', and OID:'.1.3.6.1.2.1.25.2.3.1.5.5'
06/13/2011 08:47:08 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'00', and OID:'.1.3.6.1.2.1.25.2.3.1.5.5'
06/13/2011 08:47:08 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'01', and OID:'.1.3.6.1.2.1.25.2.3.1.6.2'
06/13/2011 08:47:09 PM - CMDPHP: Poller[0] WARNING: SNMP Get Timeout for Host:'02', and OID:'.1.3.6.1.2.1.25.2.3.1.5.4'

I guess that this is the reason why sometimes cacti reads the right value and sometimes not...
But if I try snmpwalk from cacti server I don't get any timeout :-O

[noname]

At first, try to increase SNMP timeout (e.g. 500ms -> 1000ms) in your device settings.

[mgtech]

I even tried to change the "Maximum OID's Per Get Request" value from 2 to 1, but nothing has changed.
I'll try to follow you suggestions, but if it was a timeout matter, I would have the same problem with CPU Utilization, IIS statistics and network traffic... but they work correctly, as you can see in the attached file.

Anyway I'll let you know in a few hours

Thanks!

P.S.
One question: is there a way to modify all the timeout values in all the devices, without editing one by one by hand?

[noname]

I'll try to follow you suggestions, but if it was a timeout matter, I would have the same problem with CPU Utilization, IIS statistics and network traffic... but they work correctly, as you can see in the attached file.

I'm not sure, but I guess SNMP agent (snmpd) may take more time to calculate disk spaces than other type of information.
(I assumed calculating is performed by on demand..)

One question: is there a way to modify all the timeout values in all the devices, without editing one by one by hand?

1) Select hosts in devices list
2) Select "Change SNMP Options" from pull-down menu at the bottom of list
3) Click [Go]

[mgtech]

It looks like that was the issue
I changed the SNMP timeout value 2 hours ago, and since then I didn't get any more error.

Thank you very much!!!