I am still using v1.8.5 but I have not seen anything related to this problem in subsequent updates, v1.8.5 should still be relevant (updating the XML conf files of a new release takes foverer...)
Basically, I setup a simple set of thresholds based on OpenNMS examples.
Floppy will be .1.3.6.1.2.1.25.2.1.5 HDD will be .1.3.6.1.2.1.25.2.1.4 RAM will be .1.3.6.1.2.1.25.2.1.2 SWAP will be .1.3.6.1.2.1.25.2.1.3 etc.
[A] is set to filter ^\.1\.3\.6\.1\.2\.1\.25\.2\.1\.4$ and points towards diskHigh/diskRearm [B] is set to filter ^\.1\.3\.6\.1\.2\.1\.25\.2\.1\.2$ and points towards RAMHigh/RAMRearm [C] is set to filter ^\.1\.3\.6\.1\.2\.1\.25\.2\.1\.3$ and points towards swapHigh/swapRearm
Problem ======
[A] will catch ALL the threshold alerts, regardless of the filter. I get Swap/RAM/HDD alerts all sent from DiskHIGH regardless of their hrStorageType.
Additionally, I tried to filter [A] even more to avoid getting alerts from our SQL Servers for which both RAM and HDD are fully used, by design. Only way to separate these logical volumes was to filter their names (TEMP_DB, SQL_TEMPDB, SQL_DATA) in a "everything except" regexp. I added:
But nothing works. The filters are just never applied in real-time filtering.
I am the only one with this problem?
Environment
Windows 2008 R2
Acceptance / Success Criteria
None
Lucidchart Diagrams
Activity
Show:
Miskells April 25, 2011 at 4:15 AM
Also, would you mind attaching your thresholds.xml file as well?
Miskells April 25, 2011 at 4:14 AM
Could you please turn the logging on collectd up to 'Debug', and attach the logs of it failing to apply the filter properly? In particular I'm looking for some lines starting with "passedThresholdFilters:" (the bit of code that is presumably at fault here), although the context of lines around them would be useful too.
I am still using v1.8.5 but I have not seen anything related to this problem in subsequent updates, v1.8.5 should still be relevant (updating the XML conf files of a new release takes foverer...)
Basically, I setup a simple set of thresholds based on OpenNMS examples.
Group:
hrStorage C:/Program Files (x86)/OpenNMS/share/rrd/snmp/
Expression-based Thresholds
[A] high hrStorageUsed / hrStorageSize * 100.0 hrStorageIndex hrStorageDescr 80.0 70.0 1 uei.opennms.org/Thresholding/FLOW/diskHigh uei.opennms.org/Thresholding/FLOW/diskRearm
[B] high hrStorageUsed / hrStorageSize * 100.0 hrStorageIndex hrStorageDescr 75.0 65.0 1 uei.opennms.org/Thresholding/FLOW/RAMHigh uei.opennms.org/Thresholding/FLOW/RAMRearm
[C] high hrStorageUsed / hrStorageSize * 100.0 hrStorageIndex hrStorageDescr 75.0 65.0 1 uei.opennms.org/Thresholding/FLOW/swapHigh uei.opennms.org/Thresholding/FLOW/swapRearm
They all test for the same variables and the only way to discriminate between RAM, SWAP, HDD, etc. is through either: hrStorageType or hrStorageDescr
http://www.oidview.com/mibs/0/HOST-RESOURCES-TYPES.html
Floppy will be .1.3.6.1.2.1.25.2.1.5
HDD will be .1.3.6.1.2.1.25.2.1.4
RAM will be .1.3.6.1.2.1.25.2.1.2
SWAP will be .1.3.6.1.2.1.25.2.1.3
etc.
[A] is set to filter ^\.1\.3\.6\.1\.2\.1\.25\.2\.1\.4$ and points towards diskHigh/diskRearm
[B] is set to filter ^\.1\.3\.6\.1\.2\.1\.25\.2\.1\.2$ and points towards RAMHigh/RAMRearm
[C] is set to filter ^\.1\.3\.6\.1\.2\.1\.25\.2\.1\.3$ and points towards swapHigh/swapRearm
Problem
======
[A] will catch ALL the threshold alerts, regardless of the filter. I get Swap/RAM/HDD alerts all sent from DiskHIGH regardless of their hrStorageType.
Additionally, I tried to filter [A] even more to avoid getting alerts from our SQL Servers for which both RAM and HDD are fully used, by design. Only way to separate these logical volumes was to filter their names (TEMP_DB, SQL_TEMPDB, SQL_DATA) in a "everything except" regexp. I added:
^(.(?!(TEMP_DB|SQL_TEMPDB|SQL_DATA)))*$
Also tried 3 separate rules:
^(.(?!TEMP_DB))*$
^(.(?!SQL_TEMPDB))*$
^(.(?!SQL_DATA))*$
But nothing works. The filters are just never applied in real-time filtering.
I am the only one with this problem?