Software to monitor system uptime

[Jerry_109]

Hello All,

We are looking for software to monitor Operation System uptime over 30 period, and,
generate graphs to reflect this information.

Most of our systems are as follows:

HP-UX B.11.11 U 9000/800
rp7410, rp3440, rp3410

Any suggestion would be greatly appreciated.

Thanks Jerry

[Steven E. Protter]

Your sytems already monitor this.

run the command:

uptime

You can do it with cron and collect the output in a file or email.

SEP

[Jerry_109]

sorry I meant to put "over 30 day period",
and generate graphs.

[Pete Randall]

Jerry,

HP-UX is so inherently stable that I'm afraid your graph would be a flat line. I've got servers that have been up for month. Other people talk about servers that have been up for years. I would think that printing out the /etc/shutdownlog would better server your purpose.


Pete

[Victor BERRIDGE · Accepted solution]

Hi,
You could always install Big brother to generate status and graphs...
http://www.quest.com/bigbrother/

All the best
Victor

[Jerry_109]

thanks all, but this is for management. They
like pictures, even if it does show a thin line. Someone suggested measureware. Have you utilized this product to produce such results? Yes the shutdownlog can be "cut, and
pasted" but it needs to calculated.

[Steven E. Protter]

Glance is a graphical performance monitoring product. Run in X windows, it will allow you to take nice pictures.

Atternatively you can collect performance data by script and import the data into an Excel spreadsheets and acheive a breathtaking, nice to look at flat graph.

Attaching some scripts that might help collect the data.

SEP

[Paul F. Carlson]

Use a cron job to capture uptime statistics, parse it with a shell, awk or perl script, and then use gnuplot (http://www.gnuplot.info) to create a graph of your results. You can save your graph with gnuplot in .png format to give to your management.

[Scot Bean]

I am not sure MeasureWare / Glance is what you will want. It is at too low a level, showing the machine's local performance over time, the display graphs are 4 hours max.

[Geoff Wild]

Big Brother can be used - web based but not really graphical....

Make you own graphs with Measureware data..

I do this from 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 'performance report' gwild@mydomain.ca >/dev/null 2>&1



# 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


Rgds...Geoff

[Florian Heigl (new acc)]

Monitoring uptimes was quite fancy some months ago, maybe have a look at
http://www.mreriksson.net/uptimes/uptimec/ and friends.

[Geoff Wild]

Actually - I forgot to mention - if you want management to see "live" graphs, have a look at MRTG:

http://www.ee.ethz.ch/~oetiker/webtools/mrtg

I use it to monitor CPU and Memory of my servers, but you can also do uptime as well...

Here's how I monitor HP-UX with SNMP:

######################################################################


Target[sha1.internaldomain.netcpup]:.1.3.6.1.4.1.11.2.3.1.1.15.0&.1.3.6.1.4.1.11.2.3.1.1.13.0:public@sha1.p
cacorp.net + .1.3.6.1.4.1.11.2.3.1.1.16.0&.1.3.6.1.4.1.11.2.3.1.1.14.0:public@sha1.internaldomain.net
Title[sha1.internaldomain.netcpup]: CPU Load on sha1.internaldomain.net
MaxBytes[sha1.internaldomain.netcpup]: 100
Ylegend[sha1.internaldomain.netcpup]: CPU Load
ShortLegend[sha1.internaldomain.netcpup]: %
PageTop[sha1.internaldomain.netcpup]:

CPU Load on sha1.internaldomain.net

Options[sha1.internaldomain.netcpup]: integer, growright, nopercent
WithPeak[sha1.internaldomain.netcpup]: wmy
LegendO[sha1.internaldomain.netcpup]: User/System:  
LegendI[sha1.internaldomain.netcpup]: Idle/Nice:  
Legend1[sha1.internaldomain.netcpup]: Idle/Nice CPU Time
Legend2[sha1.internaldomain.netcpup]: User/System CPU Time
Legend3[sha1.internaldomain.netcpup]: Maximal Idle/Nice CPU Time
Legend4[sha1.internaldomain.netcpup]: Maximal User/System CPU Time

######################################################################

Target[sha1.internaldomain.netmem]: .1.3.6.1.4.1.11.2.3.1.1.7.0&.1.3.6.1.4.1.11.2.3.1.1.8.0:public@sha1.pca
corp.net
Title[sha1.internaldomain.netmem]: Memory on sha1.internaldomain.net
MaxBytes[sha1.internaldomain.netmem]: 100000000
Ylegend[sha1.internaldomain.netmem]: Memory Usage
ShortLegend[sha1.internaldomain.netmem]:  
PageTop[sha1.internaldomain.netmem]:

Memory on sha1.internaldomain.net

Options[sha1.internaldomain.netmem]: integer, gauge, growright, nopercent
kilo[sha1.internaldomain.netmem]: 1024
WithPeak[sha1.internaldomain.netmem]: wmy
LegendO[sha1.internaldomain.netmem]: Total:  
LegendI[sha1.internaldomain.netmem]: Free:  
Legend1[sha1.internaldomain.netmem]: Free System Memory
Legend2[sha1.internaldomain.netmem]: Total Available System Memory
Legend3[sha1.internaldomain.netmem]: Maximal Available System Memory
Legend4[sha1.internaldomain.netmem]: Maximal Free System Memory


######################################################################


Sun and others - you need to script (as snmp is limited).


# dns stats for mrtg
3,8,13,18,23,28,33,38,43,48,53,58 * * * * /usr/local/bin/mrtg-dns-stats >/dev/null 2>&1



#!/bin/sh
# generate mrtg file for dns stats
/usr/local/sbin/rndc stats
SERVER=`uname -n`
MRTG=/tmp/$SERVER.named.stats.mrtg
STATS=/etc/namedb/named.stats
#get success
tail -7 $STATS | head -1 |awk '{print $2}'>$MRTG
tail -2 $STATS | head -1 |awk '{print $2}'>>$MRTG
uptime | awk '{print $3,$4,$5}' |sed s/,//g >>$MRTG
echo $SERVER >>$MRTG
#generate mrtg file for cpu
MRTGcpu=/tmp/$SERVER.cpu.stats.mrtg
idle="`/usr/bin/vmstat 1 2 | tail -1 | awk '{print $NF}'`"
usersys=`expr 100 - $idle`

echo $idle >$MRTGcpu
echo $usersys >>$MRTGcpu
uptime | awk '{print $3,$4,$5}' |sed s/,//g >>$MRTGcpu
echo $SERVER >>$MRTGcpu

MRTGmem=/tmp/$SERVER.mem.stats.mrtg

/usr/local/bin/mem-swap4mrtg.sh mem >$MRTGmem


The rcp those files to your web server:

#------------------------------------------

Target[dnsdnsserver]: `cat /tmp/dnsserver.named.stats.mrtg`
Options[dnsdnsserver]: nopercent,growright,perhour
Title[dnsdnsserver]: DNS: Requests per hour on dnsserver
PageTop[dnsdnsserver]:

DNS: Requests per hour on dnsserver

MaxBytes[dnsdnsserver]: 1000000000
YLegend[dnsdnsserver]: reqs/hour
ShortLegend[dnsdnsserver]: per hour
LegendI[dnsdnsserver]:  Requests:
LegendO[dnsdnsserver]:  Failures:
Legend1[dnsdnsserver]: Requests per hour
Legend2[dnsdnsserver]: Failures per hour
PageFoot[dnsdnsserver]:
Note:
success is the number of successful queries the server handled.
failures is the number of queries the server received that resulted in
errors other than those covered by nxrrset and nxdomain.


#------------------------------------------

Target[dnsserver.cpu]: `cat /tmp/dnsserver.cpu.stats.mrtg`
Title[dnsserver.cpu]: CPU Load on dnsserver.pcacorp.net
MaxBytes[dnsserver.cpu]: 100
Ylegend[dnsserver.cpu]: CPU Load
ShortLegend[dnsserver.cpu]: %
PageTop[dnsserver.cpu]:

CPU Load on dnsserver

Options[dnsserver.cpu]: integer, gauge, growright, nopercent
WithPeak[dnsserver.cpu]: wmy
LegendO[dnsserver.cpu]: User/System:  
LegendI[dnsserver.cpu]: Idle/Nice:  
Legend1[dnsserver.cpu]: Idle/Nice CPU Time
Legend2[dnsserver.cpu]: User/System CPU Time
Legend3[dnsserver.cpu]: Maximal Idle/Nice CPU Time
Legend4[dnsserver.cpu]: Maximal User/System CPU Time

Target[dnsserver.mem]: `cat /tmp/dnsserver.mem.stats.mrtg`
Title[dnsserver.mem]: Memory on dnsserver.pcacorp.net
MaxBytes[dnsserver.mem]: 1000000000
Ylegend[dnsserver.mem]: Memory Usage
ShortLegend[dnsserver.mem]:  
PageTop[dnsserver.mem]:

Memory on dnsserver

Options[dnsserver.mem]: integer, gauge, growright, nopercent
kilo[dnsserver.mem]: 1024
WithPeak[dnsserver.mem]: wmy
kMG[dnsserver.mem]: k,M,G
LegendI[dnsserver.mem]: Free: 
LegendO[dnsserver.mem]: Total: 
Legend1[dnsserver.mem]: Free System Memory
Legend2[dnsserver.mem]: Total Available System Memory
Legend3[dnsserver.mem]: Maximal Free System Memory
Legend4[dnsserver.mem]: Maximal Available System Memory

Rgds...Geoff

[Ted Buis]

The package that used to graph the MeasureWare results used to be called PerfView, now renamed. It was sold separately. To get recent information contact your HP OpenView Sales Rep.

http://www.managementsoftware.hp.com/products/ovperf/index.html

also

http://www.managementsoftware.hp.com/products/ovpi/index.html