Hi ,
Is there any way of adding some more mount points in NMON Performance Monitor for Unix and Linux Systems app in "UI JFSFILE, Percentage Usage of FileSystems"?
As of now , I could see /home , /boot , /users , /var , /opt , /oem . I would like to add /swap , and some more files under /opt/webapps/*
Thanks.
Hi,
The nmon process is expected to list every mounted file system, at least at startup time (it is restarted about every 2 hours by default), as you ask the question I guess you can't find some of the file system you expect.
The easiest would be first to check using an SPL search (note you can use HOWTO interfaces):
Listing all file system:
eventtype=nmon:performance type=JFSFILE host=foo | stats count by host,device | sort limit=0 host,device
Charting over time:
eventtype=nmon:performance type=JFSFILE host=foo device=* | timechart avg(value) as pct_use by device
You should have the same list of available file system that you have within interfaces, unless there is an error I am not aware of.
In such a case it might help to force rebuilding the NMON Data JFS FILE data model that is being used by interfaces to build that list automatically.
As an advanced verification, you can directly check the content of the raw Nmon file, you will find them in $SPLUNK_HOME/var/log/nmon/var/nmon_repository/
Exemple:
grep JFSFILE,JFS /opt/splunk/var/log/nmon/var/nmon_repository/*.nmon
This will provides the first line which the csv header, and contains the list of file systems under analysis by Nmon processes.
If you can't find a file system here (such as the swap that you will not find there, it not considered as a file system and is analysed by Memory interfaces), you will not find it in interfaces of the app.
Don't hesitate to inform me the case interfaces would not provide the expected list of file systems.
Regards,
Guilhem
Hi,
The nmon process is expected to list every mounted file system, at least at startup time (it is restarted about every 2 hours by default), as you ask the question I guess you can't find some of the file system you expect.
The easiest would be first to check using an SPL search (note you can use HOWTO interfaces):
Listing all file system:
eventtype=nmon:performance type=JFSFILE host=foo | stats count by host,device | sort limit=0 host,device
Charting over time:
eventtype=nmon:performance type=JFSFILE host=foo device=* | timechart avg(value) as pct_use by device
You should have the same list of available file system that you have within interfaces, unless there is an error I am not aware of.
In such a case it might help to force rebuilding the NMON Data JFS FILE data model that is being used by interfaces to build that list automatically.
As an advanced verification, you can directly check the content of the raw Nmon file, you will find them in $SPLUNK_HOME/var/log/nmon/var/nmon_repository/
Exemple:
grep JFSFILE,JFS /opt/splunk/var/log/nmon/var/nmon_repository/*.nmon
This will provides the first line which the csv header, and contains the list of file systems under analysis by Nmon processes.
If you can't find a file system here (such as the swap that you will not find there, it not considered as a file system and is analysed by Memory interfaces), you will not find it in interfaces of the app.
Don't hesitate to inform me the case interfaces would not provide the expected list of file systems.
Regards,
Guilhem