using glance to grab a baseline

[Adam Garsha]

hello, I wish to use glance to grab a baseline of cpu/memory/disk/network utilization over this weekend.

Does anyone have a useful command-line or script all hacked out for this?

I need to be able to graph (pull into excel) this information with granularity of about 10-30 seconds per datum.

Any help appreciated. Also if there is a better tool for this, let me know.

[TwoProc · Accepted solution]

Since you've got glance, you've probably got the measure agents running and collecting data. Even without perfview, you can extract the data from the logs it builds with the "extract" command.

Do a man on extract to see what the options are.

EXTRACT(1) EXTRACT(1)

NAME
extract - reads the contents of MeasureWare Agent (MWA) scopeux log
files and creates a "extracted" file whose format has been optimized
for archiving or use by other systems for analysis.

extract - reads the contents of MWA scopeux log files or DSI log files
and exports data in ASCII, BINARY, DATAFILES or WK1 (spreadsheet)
format for use by other programs.

[Martin Johnson]

MeasureWare (MWA) or OVPA data would be better suited for this. With MWA you could use the extract utility to export the data in table format that can be loaded into Excel.

HTH
Marty

[Geoff Wild]

Here's what I do with MMA:

Run from root's cron:
# performance
55 23 * * 1-5 /opt/perf/bin/extract -xp -r /home/gwild/zgbltemplate -g -b today 7:00 -e today 18:00 -f stdout | /bin/mailx -s 'Myserver performance report' gwild@mydomain.com >/dev/null 2>&1

Here's the template
# cat /home/gwild/zgbltemplate
REPORT "MWA Export on !SYSTEM_ID"
FORMAT ASCII
HEADINGS ON
SEPARATOR="|"
SUMMARY=60
MISSING=0
DATA TYPE GLOBAL
YEAR
DATE
TIME
GBL_CPU_TOTAL_UTIL
GBL_MEM_UTIL
GBL_SWAP_SPACE_UTIL
GBL_MEM_CACHE_HIT_PCT
GBL_DISK_UTIL_PEAK
GBL_FS_SPACE_UTIL_PEAK
GBL_MEM_PAGEOUT_RATE
GBL_MEM_SWAPOUT_RATE
GBL_MEM_QUEUE


Rgds...Geoff

[Adam Garsha]

Yes, MWA is the ticket! 2 follow-ups (since I am buried in tasks, and you may know off the top of your head):

1.) To adjust frequency of data gathering, which file?

2.) Documentation? (besides man pages)? having trouble finding a good link.

[Raj D.]

Hi Adam,

Here is a script with glacne to grab the cpu/memory output , You can add the disk and network parameter with that .

Cheers,
Raj.

[Adam Garsha]

I found docs: /opt/perf/paperdocs/...

[Hazem Mahmoud_3]

The tool that we use here is called Lund Performance Solutions (www.lund.com). It allows us to keep track of cpu/disk/memory/network utilization. I just import the files it generates into Peformance Gallery Gold (an additional software that they sell) and it generates the graphs we need with baselines, time periods, etc.

-Hazem

[Geoff Wild]

Frequency - that's done from the extract command...

In my cron,

55 23 * * 1-5 /opt/perf/bin/extract -xp -r /home/gwild/zgbltemplate -g -b today 7:00 -e today 18:00 -f stdout | /bin/mailx -s 'Myserver performance report' gwild@mydomain.com >/dev/null 2>&1

I check it daily from 7AM to 6PM Mon-Fri

Change the 1-5 to a * to check every day...

you can also run it manually...

just man extract for more info.


Rgds...Geoff