About nmon command line parameters
-f spreadsheet output format [Note: default -s300 -c288] The output file is <hostname> _YYYYMMDD_HHMM.nmon
-F <filename> Same as -f, but the file name provided by the user
-c <number> number of snapshots
-d Request disk service and waiting time (DISKSERV and DISKWAIT)
-i <percent> Ignore the process to generate the TOP part when using less than this amount of CPU-useful for reducing the amount of data
-g <filename> The file containing the disk group definition
-l <dpl> The number of hdisks per sheet-the default is 150, the maximum is 250.
-m <dir> NMON change to this directory before saving the file
-r <runname> enter the spreadsheet file [default host name]
-s <second> interval between snapshots
-x capacity planning (1 day, monitoring every 15 minutes = -fdt -s900 -c96)
-t includes top-level processes in the output
-T as -t plus save command line parameters in the UARG section
-A includes asynchronous I/O (PROCAIO) part of the data
-D prevents the DISK part (useful when a disk group) is being used because there are too many hard disks to process)
-E Stop generating the ESS part (required when the disk group is in use) because there are too many vpaths to process)
-J prevents the JFS part from being generated (prevents Excel errors when you have more than 255 file systems)
-L includes the LARGEPAGE part
-N includes the NFS part
-S includes the WLM part of the tape class
-W includes WLM parts without subclasses
-Y includes the Summary part (very effective replacement -t if does not require PID level data)
If you want to manually control the generated file of nmon, you only need to set the time interval instead of the number of monitoring times, and kill the process through the linux command.
For example: ps -f | grep "nmon -f" find the nmon process
kill -9 process number kill process
If you monitor multiple servers, you can use a script, the script content is as follows:
#!/bin/sh
ps -f | grep "nmon -f"> test.txt //Export the process content
num=$(awk'NR==1{print $2}' test.txt) //Find the process number , Where NR is the number of rows, print determines the number of columns
kill -9 $num //Kill the process