I'm having an issue where esxtop generates so much data in batch mode that I cannot import it into Excel 2003 for analysis.
I'm only looking to gather some data from the disk pane and have turned off all the columns on the CPU, Memory and Network screens. I've written these changes to the .esxtop3rc file.
Even by turning off just about everything I can find, esxtop in batch mode still generates CSV files with too many columns for Excel to import.
So how do I get my esxtop into a tool what will generate some pretty graphs for me? Any suggestions?
Is there a way to configure esxtop to only capture certain counters in batch mode (a la perfmon)?
Thanks
You're limited to 256 columns in 2003. I upgraded to XL 2007 for just this reason. It's something like 16k columns in 2007.
http://blogs.msdn.com/excel/archive/2005/09/26/474258.aspx
Grab a perl or awk guru and they will be able to help you parse your data and throw out all the stuff you don't need.
Message was edited by:
Jae Ellers
Thanks.
I've actually been through the exercise of cleaning out many of the columns I don't want but even if I just include the disk fields, there are still too many columns for Excel to import.
Guess I'm off to buy my perl guy a case of beer...
http://www.perlmeme.org/tutorials/parsing_csv.html is pretty good.
http://www.google.com/search?q=perlparsingcsvORcolumn
Just open up the output of the batch mode in Windows perfmon. Within perfmon point to the CSV file output as a log-file source, and select the counters you want to view.
Perl is the best for this. We use it to parse a local csv file and expose stats over snmp.
Care to elaborate on that comment? "Perl is the best for this. We use it to parse a local csv file and expose stats over snmp."
If you know specifically what columns you want in terms of the column number, you can use the 'cut' command.
eg. to run esxtop in batch mode in the background and get it to only dump out columns 1 and 6 to a file called output.log:
esxtop -b | cut -d, -f1,6 >> output.log &
Unfortunately this doesn't help when you need certain columns based on the column header (as opposed to the column number). For instance I want to only dump out the '% Used' and '% Ready' columns..ie. all columns where the header contains the strings '% Used' or '% Ready'. That would give a much more manageable output file. I'm still trying to find out how to do that (maybe someone can can help?)