[rrd-users] Re: insert time-stamped data from a flat file?

[Gerhard Ecaroh Froehlich]

Hi Dulac

"Dulac, Daniel (Dan)" wrote:
> 
> I have a completely different tool collecting data at various time intervals
> (mostly 300 seconds) and writing this data to a flat-file.  Each row of this
> flat file contains the time the data was collected, the metric (variable) I

Why do you write in a file instead of updating instantly?

> stamp.  At the end of each day, I want to populate the RRD with this new

As far as i know you _can not_ update the whole at the end of the day. 

> data-reading from this flat-file (obviously, my script does the proper
> conversion for the RRDs::update syntax)When running my script and
> attempting to insert data into the RRD, I get the following error
> (repeatedly with different times):
> 
> ERROR: illegal attempt to update using time 1039622171 when last update time
> is 1039622235 (minimum

You want to insert data at a timestamp before the last update: 

Want to update:   > using time 1039622171
Real last update: > last update time is 1039622235

where 1039622171 is before 1039622235!

This is not possible. An update is always _past_ the last date.

> one second step)
> 
> 
> 
> Can anyone give me insight to what this error actually means?  Is it
> possible to use RRD with what I'm trying to accomplish?  I want to utilize

This would not be the intention of RRDTool to update RRDS at one time a day.
RRDTool is to update on regulary continous times step by step.

> Dan
> 

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: please help a lost newbie

[Gerhard Ecaroh Froehlich]

Hi Dan!

Dan Gahlinger wrote:
> 
> Ok I wrote a shell script to collect data from our boxes,

[...]

> How do I get this graphing the data ?
> 
> Dan.
> 

Oh, there were some good answers in the meantime. An additional hint how you can
get it quick, remember there are a lot of ways to do a thing. My sugesstion
should work without writing in a file as step between:

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

$ vi /etc/monitored_servers.conf
server1
server2
server3

then

$ vi /opt/myscripts/whos-on.sh

#!/bin/bash

HOSTLIST=/etc/monitored_servers.conf
DS=usercount
RRDTOOL=/my/path/to/rrdtool
DATADIR=/my/path/to/datadir/

for HOST in $HOSTLIST
do
USERS=`finger @$HOSTLIST | grep -c tty`
RRDTOOL update $DATADIR$HOST.rrd  $DS $USERS
done

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

and run whos-on.sh from cron every 300 seconds. Then set up a script which polls
out
the data with rrdgraph and writes a graph and html for easy viewing in a
browser. Run it from cron too.

Same loop for graphing means:

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 
#!/bin/bash
# quick sometimes pseudocode

# define your settings
HOSTLIST=/etc/monitored_servers.conf
DS=usercount
RRDTOOL=/my/path/to/rrdtool
DATADIR=/my/path/to/datadir/

# quick hack to get unixtime in seconds (needs perl, don't know it better)
NOW=`perl -e 'print time'`

#Timespan is one day (86400 seconds):
let THEN=$NOW-86400

# let the shell do the work
for HOST in $HOSTLIST
do

$RRDTOOL graph $HOST.png --start $THEN --title='Users on $HOST' \
DEF:d$DS=$HOST.rrd:$DS:AVERAGE \
LINE2:d$DS#00FF00:"Users"

done

exit 0
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 

Create one static HTML page. Done. There are lots of ways, you may glue both
togather to simplify.

Hope this is alright cause i am not a programmer. Any suggestions and
corrections are welcome.

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: 1 .cfg file or mulitple .cfg files?

[Gerhard Ecaroh Froehlich]

Hi Ed!

Ed Grant wrote:
> 
> All,
> 
> I am new to RRDTool and was wondering what is best when it comes to my
> mrtg.cfg files. I currently have only one .cfg file that is sectioned off
> per the type of monitoring I wish to do(disk, cpu, nic, etc...). What I am
> wondering is it best when using RRDTool to have multiple configs, say one
> per device being monitored or is one config file OK?
> 
> TIA
> 

More generally my response: We had MRTG before switching to RRDTool for a lot of
servers. Not only the performance but mainly the overloaded cfg file with
redundancies (by MRTG needing HTML code in the cfg file) forced us to leave MRTG
complete. There was also a need that other not that experienced administrators
have to add or remove devices. This is hazardous if someone does not know how to
distinguish HTML from configuration. If you only have a small site which less
devices stay with one. But if you expect growing numbers devide it in more cfg -
or best - leave MRTG.

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Cisco 2600 Monitoring

[Gerhard Ecaroh Froehlich]

Hi Michael,

Michael Love wrote:
> 
> I'm in the process of reading the documentation of our Cisco 2600 router...
> I want to user RRD tp mointor and graph the top ten incoming and outgoing
> ports on the router. Has anyone else on the mailing list done this already?

If i understand you right, you want to collect Ciscos Flow capablities. This
one's coverd by Ciscos Net Flow, Flow-Tools by Fulmer or Flowscan from Plonka.
You can use RRD's to store, average and graph, but to collect the flows which
will give you top ten port you need a flow collector.

<http://net.doit.wisc.edu/~plonka>
<http://www.splintered.net/sw/flow-tools/docs/flow-fanout.html>
<http://sourceforge.net/projects/ehnt>
<http://www.switch.ch/tf-tant/floma/software.html>
<http://www.slac.stanford.edu/comp/net/netflow/SLAC-Netflow.html>
<http://www.cisco.com/univercd/cc/td/doc/product/rtrmgmt/nfc/nfc_3_0/nfc_ug/index.htm>

Dear list users, correct me, if i am wrong

> ---
> Michael Love,
> SDC Technical Support
> http://www.sdc.org

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Status info for Solaris

[Gerhard Ecaroh Froehlich]

Hi Simon,

Simon wrote:
> 
> Hi,
> 
> Sometime ago I wrote a script for Linux that produces 4 graphs (CPU
> usage, Mem usage, HDA usage and Network usage).  I did this using
> /proc/loadavg, /proc/meminfo, df -k and ifconfig.  The only problem is
> that I now have to migrate this to a Sparc Solaris machine which I am
> not too familiar with.  I can still do the df -k but cannot find a
> simple way to change the script to do everything else.

Try to use the common tools like: vmstat, uptime, netstat, sar and du. Better
way to be compatible might be to switch to SNMP. This one's much bigger than
custom scripts, especialy in big environments. Remember, the small one's are
growing!  ;-)

Unfortanetly SUN Solaris SNMP Agent does not support disk usage like a "du".
This would be fixed with UCD-SNMP which i would advise you to use. It is
available for more platforms and do make you more independent.
 
For me i was to lazy (and unable) to programm it for my own. I like Remstats
which does everything for me, including monitoring Solaris without SNMP.

> Any help would be much appreciated.
> 
> Regards,
> 
> Si


Bye, Ecaroh

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Just a basic question... one or multiple RRDs?

[Gerhard Ecaroh Froehlich]

Hi Noah,

as i have many custom programmed applications which are monitored by custom
scripts from commandline tools there was a result concerning the layout of
rrd's.

In some cases you may be able to foresee that data sources may change in future.
But some are well known and always act togather. In Example:

One rrd with multiple data sources:

 - A router: has always an "in" and an "out"
 - A CPU: has "user" time, "idle" and "system" time which will result in a total

One rrd for each data source:

 - A servers harddisk(s): could be one or more which will have different
   partitions or slices.
 - A database server (oracle monitored by SNMP) which can have multiple
databases
   even for multiple customers which will be set up and delete by the time.

The thing is to group things which will be togather forever by function and can
not divided. Seperate your data sources in own rrd's if you are unsure cause
they may be deleted, added or may grow in future.

For similar data sources (i.e. you have 10 customers which get the same
monitoring) it was usefull for me to declare _all_ datasources the same name and
to differ by directoryname. Writing of scripts and copying in the desired
directory is that easy then! This solution was inspired by Thomas Erskine's
Remstats - Thanks a lot Thomas. Oh, i remember he did some good stuff with
wildcards, have a look at his documentation. This saves lots of work too.

Bye, Ecaroh

Noah Leaman wrote:
> 
> Can I create just a *single* RRD file to store all this different type
> of data? Or is it wiser to create separate RRDs... one for each
> 
> Thought I would run it by some people that experience ... this is my
> first time working with RRD.
> 
> --
> Noah

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Unix system perf monitoring/trending & RRD.

[Gerhard Ecaroh Froehlich]

Hi!

PRASAD GADGIL /INFRA/INFOTECH wrote:
> 
> hello,
> 

Okay, english is not my native language and i may have not understood everything
you meant. I try to give an answer.

> There are 2 sides to unix perf monitoring -

There are 3 sides:

 - Commandline tools like: vmstat, uptime, top, sar ...
   which are allways somekind of 'realtime'
 - Special 'manufacturer' dependend tools which are shipped with 
   the hardware like monitoring systats over embedded webservers 
   or perfmeters and so on.
 - Polling tools like SNMP
   which may be realtime if you call it from console
   but mostly are triggerd through scripts from while to while.
   This is not realtime!
 
> One soln that caters both will be great but seems not
> possible to me in my case.

You did not tell us why.
 
> impl Difficulties -
> 
> certain data has variable format (field
> positions/no) [e.g. disk r/w data due to
> variable no of discs]
> maintaining custom RRD formats for diff servers when
> these also keep changing
> from time to time is painfull (maintainance
> difficult)

You can avoid this by taking a common standard like SNMP. You may have to set ip
up seperatly, preconfigured/delivered by vendor or as free implementation like
UCD-SNMP. Most of the graphing tools like RRDtool and the wrappers round RRDtool
World rely on SNMP. It is the most comfortable way to gather data. But not the
most sure if you want to poll over the Web.
 
> Comments welcome.
> 
> regards,
> prasad gadgil.
> --

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Windows Oids

[Gerhard Ecaroh Froehlich]

Hi Serge, hi rrdusers

Serge Maandag wrote:
> 
> Dear listmembers,
> 
> I am looking for some decent OIDs to monitor a windows 2000 server.
> 
> In particular I am looking for the 5 minute cpu usage OIDs, not the 5
> minute load average OIDs.
> Also I would like to monitor the memory stats and the disk usage.
> 
> Can anyone help me in the right direction?

Please don't be angry if i missunderstood your question and intention. You can
find OIDS on:

<http://www.mibcentral.com>

therefor searching for Vendor Microsoft. This would be a lot faster than
searching MS Site. ;-))
 
> Serge Maandag.
> 
> --
> Unsubscribe mailto:[EMAIL PROTECTED]
> Helpmailto:[EMAIL PROTECTED]
> Archive http://www.ee.ethz.ch/~slist/rrd-users
> WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Error: could not lock RRD. When I am using NFS

[Gerhard Ecaroh Froehlich]

Hi,

"Saravia, Jovanny - (Col)" wrote:
> 
> Hello rrd-users.
> 
> I migrate some days ago mrtg to rrd. and everything works fine for me.
> Unafortunately I have more than 5000 targets to be monitored. I was using
> NFS in order to have 3 servers monitoring and updating in the same Web
> Server.

As long as i have read in the lists it is not possible to run RRDTool over NFS.
As a workaround you can use Apaches "rewrite engine" module. So you store all
rrds on your collecting 3 servers, setup apache on each of them and melt the
whole stuff together via your Webserver and rewrite engine. Have a look at the
apache documentation for additional information. A quick hint is on the end of
this mail.

If you plan to back things up, try to use mirroring. At least, i know mirroring
is NO backup. Additionaly new Sun Solaris comes out with a brand new feature to
take 'snapshots' like you may know it from NetAppFiler. 

> Rgds,
> 
>     Jovanny Saravia

bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator 


--- Apache Rewrite Engine (module required) ---

RewriteEngine on
RewriteRule   ^/collectorhost1/(.*) 
http://collectorhost1:1234/collectorhost1/$1  [P]
RewriteRule   ^/collectorhost2/(.*) 
http://collectorhost2:1234/collectorhost2/$1  [P]
RewriteRule   ^/collectorhost3/(.*) 
http://collectorhost3:1234/collectorhost3/$1  [P]

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Compiling of rrd-tool

[Gerhard Ecaroh Froehlich]

Hello Carsten,

[EMAIL PROTECTED] wrote:
> 
> Hi.
> 
> I'm new to rrd-tools, and I'm having some problems in compiling it.
> I'm working on a SGI Unix machine with IRIX 6.5.16m as OS.
> Using perl 5.004, gcc 3.04 and rrd-tool.1.0.39.

Look at <http://people.ee.ethz.ch/~oetiker/webtools/rrdtool/compiling.html>. You
need perl 5.005 to compile correct. Perl 5.004 will _not_ fit, i tried it too
cause of lacking perl 5.005 and did not succed on Solaris. Install perl 5.005
first and start again.

bye, Ecaroh

> 
> Regards
> Carsten Soetofte
> Sys Admin
> 
> --

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: How to monitor a server

[Gerhard Ecaroh Froehlich]

Hello,

in short: It is possible. On the Win side you have to install snmp service on
NT4 and W2K. On NT4 you have to reinstall the service pack and to copy snmp*.dll
(i do not remember exactly the file name) back cause this one gets destroyed by
installing snmp services. This is described in Knowledge Base. Search there for
snmp.dll and snmp-services if you need additionell help.

For the MRTG site of setup have a look at the documentation on:

<http://people.ee.ethz.ch/~oetiker/webtools/mrtg/>

As small annotation of my unimportant oppinon: Is is a lot easier to install
MRTG on Linux.


Bye, Ecaroh


"Luquin, Andres" wrote:
> 
> Hi everybody:
> 
> I'd like to monitor the CPU load and traffic in several servers
> (with NT4 and W2K) using MRTG. I don't know if it's possible. Thanks a lot
> in advance !!! Cheers.
> 
>             Andrés.

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: problem building a particular mrtg.cfg

[Gerhard Ecaroh Froehlich]

Anup Ray wrote:
> 
> Hi,
> I am getting a problem while making mrtg.cfg file for a solaris box.
> Following is the error output:
> 
> --base: Get Device Info on [EMAIL PROTECTED]:
> SNMP Error:
> Received SNMP response with error code
>   error status: noSuchName
>   index 1 (OID: 1.3.6.1.2.1.1.1.0)
> SNMPv1_Session (remote host: "u10-1.safe-prop.com" [10.10.1.100].161)
>   community: "public"
>  request ID: -1281356335
> PDU bufsize: 8000 bytes
> timeout: 2s
> retries: 5
> backoff: 1)
> But when I use it for other systems (solaris) it works. I am using command
> from a single system(ultra30). So same source programs are being used.

Try to run the SNMP query (for instance with UCD-SNMP with "snmpwalk") from the
commandline. You should see the following string. If not, snmpdx doesnot run,
the firewall rules do not pass querys on Port 161, the community string is
wrong, the snmpd is not configured.

host:admin,94% ./snmpwalk -p 161 host public .1.3.6.1.2.1.1.1.0
system.sysDescr.0 = Sun SNMP Agent, UltraSPARC-IIi-cEngine

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: problem building a particular mrtg.cfg

[Gerhard Ecaroh Froehlich]

Anup Ray wrote:
> 
> Thank you Dom for your suggestion. Configuration was pkay. I just restarted
> snmpd daemon and it started working okay.
> Anup

And just to add a hint for later production: Solaris snmpdx *likes* to hang
sometimes. And on Solaris 2.6 network speed is reported wrong. If you get
stopping graphs on your machine, try the following:

[EMAIL PROTECTED]> /etc/init.d/init.dmi stop
[EMAIL PROTECTED]> /etc/init.d/init.snmpdx stop
[EMAIL PROTECTED]> /etc/init.d/init.snmpd start
[EMAIL PROTECTED]> /etc/init.d/init.dmi start

This helps. Or if you paranoid make an script to run it from cron in the night.
I have to restart it from while to while on my solaris hosts.

Ecaroh

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: rrdtool-1.0.39 on Solaris 6 : configure error

[Gerhard Ecaroh Froehlich]

Hi Jerome,

"Boismartel, Jerome" wrote:
> 
> Hello,
> 
> When I run sh configure on my system I have the following error :
> checking for the C compiler perl wants to use to build its modules...
> ./configure: test: unknown operator -B/usr/ccs/bin/
> 
> I run Solaris 6 and perl 5.8.0
> 
> Does anyone run into this issue ? What was the solution ?

One solution is to read the archives. The question was asked multiple only weeks
ago. In short: Install all GNU Tools and GCC, including correct settings of
$PATH. This should help, cause Solaris don't compile that good.

> ==
> Jerome Boismartel


Ecaroh

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: measure the total traffic

[Gerhard Ecaroh Froehlich]

Hi T. Xar!

"T. Xar" wrote:
> 
> Hi

[ snip to much quotings ;-) ]

> Now i decided to log also the sum of traffic and can show the monthly result
> the ISP.

[ snip to much quotings ;-) ]

> It should count the sum of the in and out going traffic of a IP range or one
> IP or only a web server, etc..
> I have hoped that the combination of snmp and rrdtool can do so.

It can do sums of traffic, this is total amount what passes over a certain NIC
or router (if the device is capable of snmp). This can be setup quiet easy and
quick with MRTG, 14all and RRDTool. If you do not have snmp and use unix
machines you can use "netstat" and some custom scripts.

> So which tools are using the many ISPs outthere to charge (corrrectly) with
> their customers ?

What you want is monitoring on a "service" base too. This is not covered by
normal snmp nor rrdtool alone. You can log this on a "flow based" view of
traffic, which is commonly a feature of Ciscos IOS, called "NetFlow". A Cisco
Router (and some others which i do not remember) can stream out a "flow" which
has accounting information of bytes, packets and ports which you can use to
track down network traffic on a service based summing.

<http://www.cisco.com/warp/public/732/Tech/nmp/netflow/index.shtml>
<http://www.caida.org/tools/utilities/>
<http://www.caida.org/tools/utilities/flowscan>
<http://www.splintered.net/sw/flow-tools/docs/flow-fanout.html>
<http://net.doit.wisc.edu/~plonka/FlowScan/#Announcement>
 
> Thank you in advance
> Thomas Xar

And at least, please be nice to us and shorten your quotations. Thanks.


Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: MRTG and Apache

[Gerhard Ecaroh Froehlich]

Hello,

Thilanka Samarasekera wrote:
> 
> Is there anyway to monitor an apache server in windows
> using MRTG or RRD? If so where can I read about it?
> Thank you all.
> Thilanka

Yes. I am not familar with the different front and backends round rrdtool, but
you can a) write your own perl script to extract hit counts, byte counts out of
your access_log of apache or b) use a tool like remstats, who has this
functionalty build in.

But be aware that is more difficult to extract data from logs than querying an
snmp device. And it may be more difficult to set things up on windows than on
unix. ;-)

Bye, Ecaroh

<http://remstats.sourceforge.net>
 


-- 
Gerhard Ecaroh Froehlich, Systemadministrator 
Real Time Investmentbanking (Team Breitenversorgung)
Telefon: +49 (0) 69 7447 5117
Geno-Rechenzentrum GmbH, 60528 Frankfurt / Main, Hahnstrasse 46

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Values in RRD Archive

[Gerhard Ecaroh Froehlich]

Hi too!

Shyam wrote:
> 
> Hi,
> 
> Is it possible to view the values of the variables in the RRA ? If I could, 
> it would help me in confirming that the values are indeed getting updated by 
> my "update" command. ( I have scheduled my "rrdtool update" command to run 
> every minute ).

 $ rrdtool dump myrrd.rrd > myrrd.xml && more myrrd.xml

for an xml dump with the values of the rrd

 $ rrdtool info myrrd.rrd | more

for an info what kind of DS is in an rrd

 $ rrdtool graph [ .. parameters .. ]

for using the plotting features. But you have to specify what you want. Read the
man pages that come along with rrdtool. The use is well documented on rrdtool
website in the tutorials.
 
> Thanks in advance.
> 
> Regards,
> Shyam


bye, Ecaroh

--

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: rrdtool-1.0.35

[Gerhard Ecaroh Froehlich]

Hello,

Harry Nguyen wrote:
> 
> I was trying to install the rrd-tool-1.0.35 in my solaris 2.7. I untar the 
> tar ball, and tried
> to follow the instruction on README file to install it. When I did the 'sh 
> configure',
[...]
> checking for the C compiler perl wants to use to build its modules... 
> configure: test: unknown operator -B/usr/ccs/bin/
> 
> What did I do wrong!

Probably - and often questioned - Solaris do not have GNU Tools. Compiling fails
if you use build in Solaris tools like nroff, or cc instead of GCC. Change this,
install GNU software, check for $PATH when calling "sh  ./configure". This
should work. An adress for precompiled or usefull software round solaris is
<http://www.sunfreeware.com>. Look for the right Solaris version (2.7, 2.8 ...).

My system:

admin-1:ecaroh,10% pwd
/opt/gnu/gcc-2.95.3/bin
admin-1:ecaroh,11% ls 
c++  chill  g++  gcc  gcjh  jcf-dump  protoize  unprotoize
c++filt  cppg77  gcj  gcov  jv-scan   sparc-sun-solaris2.6-gcc
admin-1:ecaroh,12%

I have installed gcc and all necessary tools and libraries, it works fine.

> Thanks


Bye, Ecaroh


Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Your Compiler does not do propper IEEE math

[Gerhard Ecaroh Froehlich]

Hi,

Gerald Wilson P wrote:
> 
> Dear RRDTOOL Users,
> 
> I am having HP-9000 with HP-UX 11i
> I am facing problem during sh configure
> 
> checking if IEEE math works with signal(SIGFPE,SIG_IGN)... no
> configure: error:
> Your Compiler does not do propper IEEE math ... Please find out how to
> make IEEE math work with your compiler and let me know
> ([EMAIL PROTECTED]).
> Check config.log to see what went wrong ...
> 
> Please help me out.
> 
> Regds
> Wilson

You did not tell us, which compiler you used. I guess you did not have Gnu C
Compiler (gcc). Almost all GNU and Open Source Tools rely on gcc, which indeed
does better work as the preinstalled of traditional unix os like aix or solaris.
Also append config.log.

Bye, Ecaroh

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Counter reset to zero initialy and over hours, no rrd updates

[Gerhard Ecaroh Froehlich]

Hello,

i am monitoring Oracle DB. Unfortunately counters will be initialy "zero" when
creating the database and gets reset every night to "zero" by restarting the
databases too. For hours there will be no user request (04:00 - 09:00), therefor
after reset to "zero" the snmp value will be "zero" till the first user request
comes in the morning.

In the below shown listing i have a snmp-ora-cmr-scms.rrd who was created, but
never got an update unlike snmp-ora-chr-scms.rrd which works fine. "snmpwalk"
for the updated rrd results in real value, for the not updated i get a "zero"
cause this *is* the real *good* value which may only will increase if theres a
bad configuration on oracle db.

So i have NaN#S and UNKN for a value, which normaly indicates a correct
configured system. Do i have the wrong datatype "COUNTER"? For Oracle OID in
this case it should be Counter32.

Anybody knows how to deal with long time "zero" and initialy "zero" counters?
How can i avoid having NaN caused by zeros. 

Another question: By the nature of the oracle answer some values can be "zero".
For instance, a Ratio could result in a timespan of values like "calls /
commits", here maybe "23456 / 0". But dividing through zero is no good idea.
What can i do in this case?

Any help is appreciated.


bye, Ecaroh


--- Listings ---

# State of rrds
monhost:root,29% date
Fri Aug 23 12:10:51 MEST 2002
monhost:root,30% ls -ltr
total 926
-rw-r--r--   1 root bin  45800 Aug 23 12:08 snmp-ora-chr-scms.rrd
-rw-r--r--   1 root bin  30768 Aug 22 17:40 snmp-ora-cmr-scms.rrd

# Command for target snmp-ora-chr-scms.rrd
monhost:root,31% ./snmpwalk -p 161 oradb secret .1.3.6.1.4.1.111.4.1.1.1.2.2
enterprises.111.4.1.1.1.2.2 = Counter32: 43716


# Command for target snmp-ora-cmr-scms.rrd
monhost:root,32% ./snmpwalk -p 161 oradb secret .1.3.6.1.4.1.111.4.1.4.1.5.10.6
enterprises.111.4.1.4.1.5.10.6 = Counter32: 0
# The returnvalue "0" is the expected and wished *good* value!

# Dump for snmp-ora-cmr-scms.rrd


 0001 
 300  
 1030030517  
 UNKN 
 0.00e+00 
 17 



 lcp 
 COUNTER 
 600 
 0.00e+00 
 NaN 


 UNKN 
 0.00e+00 
 17 




 AVERAGE 
 1  
 1.00e-01 


 NaN0 

--

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: help - no graphs or html pages generated

[Gerhard Ecaroh Froehlich]

Hi "Ringler"!

You should have been reading manuals. After MRTG migrating to logformat RRDTool
MRTG does not graph anymore. This what is written in the docs. You have to write
your own Tool which uses "rrdtool graph" to generate Graphs and some cgi code
which will deliver it via Webserver and HTML. Or you look at the rrdtool
homepage, under rrdworld where you can find some prewritten software, which will
be able to feed and extract a rrd.

Bye, Ecaroh




"Ringler, Scott" wrote:
> 
> Hi all,
> 
> I have been running MRTG for about 2 weeks and think it great.  I am now
> trying to convert the graphing part to RRDTOOL with no luck.  Below is a
> sample on the cfg file that I am trying to use with rrdtool, what is
> happening is I am getting the .rrd log files in the work directory, but no
> graphs or html pages.  I can not seem to find any documentation that seems
> to help me.  Any suggestions would be greatly appreciated.
> 
> Thanks,
> Scott

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: rrdtool-1.0.39 compiling error

[Gerhard Ecaroh Froehlich]

Hi too,

seems you didn't setup $PATH correctly. You use Solaris "cc" which is always a
mess. Try to link or change $PATH to use "GNU cc". On my Sun Sparc Blade with
Solaris 8 and mounted GNU Tools under tcsh you can see clearly the diffencies
when calling Sun "cc" or GNU "cc":

myhost:froehl,1% where cc
/opt/gnu/bin/cc
/usr/ucb/cc
myhost:froehl,2% /opt/gnu/bin/cc -v
Reading specs from /opt/gnu/lib/gcc-lib/sparc-sun-solaris2.6/2.7.2.3/specs
gcc version 2.7.2.3
myhost:froehl,3%  /usr/ucb/cc  -v
/usr/ucb/cc:  language optional software package not installed
myhost:froehl,4% 

And this mostly is responsible for failures under Solaris. And you can have
those problems under Solaris with Perl too. Watch for the right versions, $PATH,
@INC and so on. Solaris is quiet different to Linux on which most OpenSource
compiles pretty good.

I am going to holidays for 2 weeks, answers to questions will be after my
return.

Bye, Ecaroh


[EMAIL PROTECTED] wrote:
> 
> Hi,
[...]
>  -KPIC -I/usr/perl5/5.00503/sun4-solaris/CORE -DPERLPATCHLEVEL=5 RRDs.c
> 
>  /usr/ucb/cc: language optional software package not installed
> 
>  *** Error code 1
> 
>  make: Fatal error: Command failed for target `RRDs.o'
> 
> 
>  Regards,
> KZ
> 
> --
> Unsubscribe mailto:[EMAIL PROTECTED]
> Helpmailto:[EMAIL PROTECTED]
> Archive http://www.ee.ethz.ch/~slist/rrd-users
> WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: rrd files not written

[Gerhard Ecaroh Froehlich]

Hi,

i am not familar with windows, but in formerly written discussions there's
always the hint to "double" backslashes under MS-Windows. Under Unix the
"backslash" stands for "IGNORE the special meaning of the following character".
So in your case check by example:

D:\\applications\\denika\\data

in all your setup. This means that RRDTool handles your setup without double
backslashing in:

D:applicationsdenkiadata

So, this is not your path. ;-)

Bye, Ecaroh


johan wrote:
> 
> Hi,
> 
> I am using rrd tool combinad with mrtg on windows 2000 server based on
> routers2.cgi. For some strange reason, he doesn't want to created the
> rrd-files. Nevertheless, all rights are set correct (cgi is writable,
> directory's for rrd-logs are set fine, etc...). I receive the following
> errors :
> 
> Error reading RRD database
> D:\applications\denika\data\rrd-logs\213.193.199.2_3.rrd
> opening 'D:\applications\denika\data\rrd-logs\213.193.199.2_3.rrd': No
> such file or directory
> Target
> RRD FileModeIn MenuIn SummaryIn/OutIn CompactNotes
>  213.193.199.2_7
> D:\applications\denika\data\rrd-logs\213.193.199.2_7.rrd
> Unable to read RRD file!
> 
> Any idea what I do wrong ?
> 
> Greetz,
> Johan
> 
> --
> Unsubscribe mailto:[EMAIL PROTECTED]
> Help    mailto:[EMAIL PROTECTED]
> Archive http://www.ee.ethz.ch/~slist/rrd-users
> WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

-- 
Gerhard Ecaroh Froehlich, Systemadministrator 
Real Time Investmentbanking (Team Breitenversorgung)
Telefon: +49 (0) 69 7447 5117
Geno-Rechenzentrum GmbH, 60528 Frankfurt / Main, Hahnstrasse 46

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Convert from MRTG to RRD

[Gerhard Ecaroh Froehlich]

Hi,

have a look at:

<http://people.ee.ethz.ch/~oetiker/webtools/rrdtool/rrdworld/index.html>

this would be good starting point. There's a lot of "open source" written code
on which you can build your environment. Read the archives, there are some
recent postings round front and backends with personal annotations and reports
of useability.

As a quick hint from my experience: Do not use MRTG as a collector for RRDTool.
MRTG is a good piece of software for small environments but is not that
extendible nor has it compact configurations files. Or do you want to end up in
twice x 10 000 lines "mrtg.cfg" like me before migrating to rrdtool?

Bye, Ecaroh


Giovanni Enriquez Gutierrez wrote:
> 
> Hi.
> 
> I'm introducing to the RRd's world, I would like to know more about RRDtool
> to be able to do my own database and feed this with RRDtool. I'd like to
[...] 
> Giovanni Enriquez
> ICQ: 8266266
> 
> > Hi all -
> >
> > Can someone offer some help or sample code?
> >
> > Thanks - Max


-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: advice needed..

[Gerhard Ecaroh Froehlich]

Hi Ross!

Ross Tsolakidis wrote:
> 
> Hi there,
> 
> monitor any/every port on that switch.  Those familiar with Cisco switches
> would know what I mean.
> Be able to calculate the usage of that box.  Be able to exclude certain
> Break down/graph different types of traffic, such as ftp, http, etc..
> Break down/graph highest transfers from and to which IP, etc..
> --
> Ross.

This seems to be a little more complicated than rrdtool/mrtg. This sort of
accounting is extraction of flows. Known as Cisco's NetFlow Feature. Most of
Cisco routers can do netflow accounting. For the collecting of this flows you
need additional software. A well known name is Dave Plonka's Netflow. Have a
look at this links:

<http://www.cisco.com/warp/public/732/Tech/nmp/netflow/index.shtml>
<http://net.doit.wisc.edu/~plonka/FlowScan/>
<http:///www.caida.org/tools/measurement/cflowd/>

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: mrtg-rrd page refreshing

[Gerhard Ecaroh Froehlich]

Hi Jay,

Jay Borkenhagen wrote:
> 
> Hi RRD Users,
> 
> My problem is that after loading a page, it soon refreshes with just
> one of the component PNGs.  For instance, after loading

This first depends on the caching mechanism of your browser. Web Writers do know
that problem too. You could set your preferences to "always look for new
version" in your browser cache. In Netscape you do have to "SHIFT + KLICK
RELOAD" to force Netscape to reload the entire webpage new. Unfortunately this
also does not work everytime. Return to "Home Page", do clear browser memory and
disk cache, exit Netscape and start it again. Painfull, indeed!

> Instead, I would like the PNGs to update at appropriate intervals and
> be displayed within the eth0.html context, like what plain MRTG would
> do.

The second place is apache. This behaviour is controlled by the configuration of
your webserver. For instance in httpd.conf of Apache you need:

# example of the "mrtg" part of httpd.conf

  ExpiresActive On
  ExpiresByType image/gif "now"
  ExpiresDefault "access 10 seconds"

  
 ExpiresActive On  # enable expirations
 # five minutes
 ExpiresDefault M300
  

  
 ExpiresActive On
 ExpiresDefault M1800
  

  
 ExpiresActive On
 ExpiresDefault M7200
  

  
 ExpiresActive On
 ExpiresDefault M86400
  

  
 ExpiresActive On
 ExpiresDefault M300
  

  # index.html is not automatically generated
  
 ExpiresActive Off
  

  
 ExpiresActive Off
  


If your not common with apache read the docs or ask your internet team.
 
> I do sometimes need to watch my MRTG in real time, to make sure that
> changes I make result in the desired traffic adjustments, so getting
> this fixed is pretty important to me.

If you want it "real time" you have to use commands on the shell. Every update
in rrdtool needs its time to appear. Commonly if you sample in 5 minute
intervalls it could happen that you need 2 runs (10 min.) until changes come
visible. Thats far from "real time". At last rrdtool is no real time tool but a
long time graphing tool.
 
> Thanks for any help.
>         Jay B.
> 
> --
>   Jay Borkenhagen [EMAIL PROTECTED]
> 

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Install problem

[Gerhard Ecaroh Froehlich]

Hi,

Leonard Miller wrote:
> 
> I do have a /usr/lib/perl5/5.6.0 directory, but Config.pm resides in
> /usr/lib/perl5/5.6.1/i386-linux.
> 
> How can I tell RRD to look at the other directory?

Link it or copy it. And try:

$ perl -e 'print @INC, "\n"'

to show your perl environment.

At least install Config.pm new. But if you have different perl versions on your
system things could get disorderd. So try also:

$ tcsh
$ where perl

or if you want to be *really* shure

$ find / -name 'perl' -exec ls -la {} \;

to look for additional perl's on your system. Make sure your path will cover the
right perl version. Different perl's have different environments in "@INC".
 
> Thanks in advance.
> Leonard

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: RRD -- front-end ?

[Gerhard Ecaroh Froehlich]

Hi,

"Vicky O. Mair" wrote:
> 
> hi there,
> 
> i am currently reading through faqs, mailing list and anything and
> everything that i can put my hands on as far as rrd goes so please bare
> with me while i try and digest the information.but in the meanwhile
> just wondering if someone could help answer a simple question.

There is no simple answer to your simple question. You did not define how large
your environment is, nor where it may end. Neither what kind of processes you
want to monitor. A brief description as short list could have make it more clear

> what are people (in general) using as far as the front-end goes for rrdtool
> database. just trying to understand and looking for recommendation.

To short it for all here: My prefered tool after long time of testing is
Remstats. The Reason to use Remstats was it's scalabilty, clear structed code,
written in pure Perl and ease of administration. And it runs without adding
additional software and without to much trouble on the beasty Solaris. :-|
 
> any help will be appreciated.
> 
> cheers,
> /vicky
> 

To those who already have a solution, the rest of this reply maybe to long and
boaring for you - go back to your work! ;-))

I don'nt know how long you'r in thist list. The questions was answered sometimes
before. To sum it up here are some excerpts of previous written mails:

I have searched long time for a usefull tool for migrating from MRTG to RRDTool.
Mainly cause i am no programmer at all. At least i have a few links collected
which i can comment after a period of testing.

Tried almost all. Took a lot of time - whow - and got lots of errors! :-(( . But
to be fair, mostly errors were caused by my cruel environment of Solaris Hosts
and
not of the used software.

An very easy way is to still let mrtg do the collecting and add 'use rrdtool' in
mrtg.cfg. This is described in the MRTG/RRDTool Documentation. As data is now
stored in rrdtool and no html and pictures are generated by mrtg anymore you
need to extract it by an cgi. Usefull in this case:

 - 14all.cgi, <http://www.uni-ulm.de/~rbawidam/mrtg-rrd/>
   Usefull, fast to setup but it doesnot scale well in big environents
 - cgi Scripts <http://www.cheshire.demon.co.uk/pub/>
   Takes automaticaly his setup from your existing mrtg.cfg without to much
   configuration. Scales good up to 20 hosts but i had some difficulties 
   to change defaults. Remember, i am not a programmer ;-(
 - cricket, i didnot test this. My first impression was that it does fit 
   network devices better than host performance. But i can be wrong.

For larger environments there are two solutions which use rrdtool without mrtg.
Not all old archives can be migrated:

 - NRG <http://nrg.hep.wisc.edu>
   Seems to work fine on Linux, my Solaris brought some difficulties
 - Remstats <http://remstats.crc.ca>
   Good for very large environments

Remstats is my favorite tool cause it's pure perl and cgi. It does not need to
much special software, modules or libraries nor does it need binaries. It's
quick to set up in an closed environment where you don't have the permission to
change existing "jumpstarts" at all. Remstats is quiet good if you have a lot of
machines ( > 50 hosts ).

You can also try Big Sister, Big Brother, Orca and so on. But if you only need
graphing as replacement for MRTG with some kind of eventnotification it will be
best. One tricky feature on Remstats is the documentation, which will be created
real time when compiling with your "configure prefixes".

Without some knowledge round compiling, networking, TCP/IP and SNMP it would be
not simple to set it up. RRDTool is no clicking tool but a scalable flexibel
tool for people who have knowledge in scripting. And setup in large environments
( > than 100 hosts) would be a project for months to do. And after all you will
see that you *have* to test a few of them to find to one that fit _your_ needs
really!

So this is my _personal_ opinion. I do not want to say, that all those other
tools are worse but that it depends on your needs and your *knowledge* what you
take. ;-))  If i would be able to programm my own tool surely my decision could
be different. But without reinventing the wheel Open Source makes you able to
customize and extend given code to your needs. And this matches all tools round
RRD World!

Sure there are a lot of other good tools round rrdtool and networkmanagment.
Some need lots of experience, some are quiet easy to install. Try it out. If you
have a Unix System most things would be easier, best is Linux at all cause most
of required modules, scripts, compilers are availabel for Linux. I had to set it
up on Solaris - this one is really hard work sometimes. Maybe someone in the
list will reply with additional information.

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mail

[rrd-users] Re: CDEF, RPN and multiple calculations using paranthesis

[Gerhard Ecaroh Froehlich]

Hi Luc

"Luc Brouard (mailing lists)" wrote:
> 

> > So to follow the tutorial with the stacks:
> >
> > Step: 1|2|3|4|5|6|7|8|9| 10|11
> > -|-|-|-|-|-|-|-|-|---|--
> > RPN:  a,b,+,c,-,a,b,+,/,100,*
> >
> > 0 ||
> > 1   |3314928|
> > 2   |3314928|263352|
> > 3   |3314928|263352|+|
> > 4   |3578280|173783|
> > 5   |3578280|173783|-|
> > 6   |3404497|3314928|
> > 7   |3404497|3314928|263352|
> > 8   |3404497|173783|263352|+| # Processing with 3 in stack?
> > 9   |3667849|437135|/|

What i do not understand is where to change the operational results in the
stack. Where does processing go on line 8? If i put "+" on the stack, which
value is changed by that? Does "+" means (for processing line 8) "263352 +
3667849", ""263352 + 173783" or something else?

> > 8   |3404497|173783|263352|+|
^--^ add this ?
/
v---  there ? 
> > 9   |3667849|437135|/|

I wonder how calculations flow while having multiple values on a stack.
 
> Why between 8 and 9 the first number changes ? (3404497 to 3667849),

Cause i asumed 3 operand on stack 8 will added to the first one.

> If this works like this then you are right to work around but I think it
> should not be needed.
> 
> Luc

I try this now for netappcpu. But unfortantely the results take 10 Minutes to
see for every test. I should write a script to work interactive.

Thanks, maybe you can explain my above question more in detail. Bye till
tomorow, work's over.

Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] CDEF, RPN and multiple calculations using paranthesis

[Gerhard Ecaroh Froehlich]

Hi,

i have some question on cdef, did not find it in the docs.

I have to sum up Oracle's cachehitratio. To get an result in percent i have to
calculate (the real world values) from snmpget:

a = consistentgets = 3314928
b = blockgets = 263352
c = phyreads = 173783

( ( 3314928 + 263352 - 173783 ) / (  3314928 + 263352 ) ) * 100 = 95 

(simplified):

 ( a + b - c ) / ( a + b ) * 100 = RESULT IN PERCENT


So to follow the tutorial with the stacks:

Step:   1|2|3|4|5|6|7|8|9| 10|11
-|-|-|-|-|-|-|-|-|---|--
RPN:a,b,+,c,-,a,b,+,/,100,*

0   ||
1   |3314928|   
2   |3314928|263352|
3   |3314928|263352|+|
4   |3578280|173783|
5   |3578280|173783|-|
6   |3404497|3314928|
7   |3404497|3314928|263352|
8   |3404497|173783|263352|+|   # Processing with 3 in stack?
9   |3667849|437135|/|
10  |8.39|100|  # which should be wrong
11  |8.39|100|*
12  |839|   # PERCENT

This is not the expected 95%.

May be it is not possible to paranthesis this way. Am i right to work around:

CDEF:cal1=a,b,+,c,-,
CDEF:cal2=a,b,+
CDEF:fin=cal1,cal2,/,100,*

Result: "fin=95"

Surely i do not fully understand the capabilities of rpn. Help is greatly
appreacheated, thanks.

bye, Ecaroh


-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Up time percent

[Gerhard Ecaroh Froehlich]

Hi all,

Olivier Nicole wrote:
> 
> I think those information can be retreived directly with SNMP, cant
> they?
> 
> Olivier

Only in short, your quiet right. Didn't think to much on it cause "Remstats" got
it all in. You may have a look at Remstats sourcecode and documentation. Here's
an example picture of my setup of Remstats.

The picture shows uptime in days as green line. You see growing uptime and
stoppings by reboot or no line for no availabilty of snmp. Sorry that i do not
apend code for that, Thomas Erskine write it: Thanks to Thomas!


bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

-- Attached file removed by Listar and put at URL below --
-- Type: application/octet-stream
-- Size: 4k (5088 bytes)
-- URL : http://www.ee.ethz.ch/~slist/pantomime/snmpuptime-3month.gif


--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] How to differ spikes from real traffic

[Gerhard Ecaroh Froehlich]

Hi,

i've set up RRDTool (SNMP) and everything works fine. So me too has some spikes
in the graphs. But most of them are looking like there just heavy traffic when
copying large files or databases. So how can i differ short heavy network
traffic from spikes? Should i imagine a real spike as "single" line up or can
there be 2,3 or 4 upgoing lines around? Are there other ways to differ real
traffic from sudden spikes?

I feel this is difficult cause our network is pretty good and consists mainly of
routers and switches and therefor is capable of transfering up to 100 MBit
(Gigabit lines too). Interfaces are allmost all SunQuad NIC. And it's not easy
to track down whether a cron made a backup or it's just another machine in that
bunch of hundreds of servers doing something other.

I.E.:

Could be short heavy traffic?

   |
   |
   |
   |
   |
___ __-|-|-_--__
Max: 75M Avg: 600 kbit


Maybe spike?

  |
 ||
 ||
 ||
 ||
 -___||___-_
Max: 75M Avg: 300 kbit


Real traffic as known:

  |
  |
  ||
  |   ||   |
  ||  ||  ||
 __|||_-__--__||___-____
Max: 10M Avg: 2M



Thanks you all for help.


    Ecaroh


-- 
Gerhard Ecaroh Froehlich, Systemadministrator

-- Attached file removed by Listar and put at URL below --
-- Type: application/octet-stream
-- Size: 268k (275182 bytes)
-- URL : http://www.ee.ethz.ch/~slist/pantomime/spike-or-not.png


--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] OID for Sun Solaris system load and disk usage

[Gerhard Ecaroh Froehlich]

Hello and excuse me,

i know this is not the right place on SNMP Discussions. I searched the 
web for over than an a day but found no hint. I monitor Sun Sparc with 
Solaris over SNMP with RRDTool and want to graph disk usage and system
load. 
(Not cpu load!). Same as 'uptime' and 'df -k' will show. Remstats is
capable
to do this with its unix-status-collector but for some reasons this not
a
suitable solution. Thanks for reply if you can help me. Maybe with an
link 
or the oid. Maybe (and i think so) it is not implemented :-(( , if this
is 
right and you know sure tell me that i should stop searching.

Thanks a lot, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: avaibility & response time

[Gerhard Ecaroh Froehlich]

Hi Amiel

David Amiel wrote:
> 
> Hi all,
> 
> I would like to monitor avaibility rate et response time for several
> services (http,pop,smtp and imap), and log results in rrd.

Try <http://remstats.crc.ca>. Alternativly there are quiet a lot of
links on the main RRDTool Website under RRD World.
 
> David
> 

Bye, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Graphing over NFS

[Gerhard Ecaroh Froehlich]

Hi RRD users,

Alex van den Bogaerdt wrote:
> 
> rrd wrote:
> 
> Perhaps you could generate the .png files locally and serve
> those over nfs?
 
Or you include 'webserver-rrd' in 'webserver-real' Apache Configuration.
So one webserver can include the other. But this more Apache specific.
Look at httpd.apache.org.

For example on 'webserver-real' (The one who servs to all admins):
--- start piece of httpd.conf ---
RewriteEngine on
RewriteRule   ^/webserver-rrd/(.*) 
http://webserver-rrd:8080/webserver-rrd/$1  [P]
--- end piece of httpd.conf ---

And on 'webserver-rrd' (The one on which rrd runs):

 a normal setup of httpd.conf


Maybe this wasn't really what you want - but maybe it's an possibilty. I
am not an apache administrator but there should be quiet a lot of
options to configure.

cheers, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: SNMP - what do i do wrong?

[Gerhard Ecaroh Froehlich]

Hi spiekey!

spiekey wrote:
> 
> Hello!

[...]
 
> Then i get this:
> # snmpget 192.168.1.40 read_community_name system.sysDescr.0
> Timeout: No Response from 192.168.1.40.
> 
> ;(
> Spiekey

I.e. UCD-SNMP try to get the syntax right in order and take full OID's:

Usage: snmpwalk [options...]  {} []

Usage: snmpget [-Cf] [options...]  {} [
...]

At least try:

# snmpget 192.168.1.40 read_community_name system .1.3.6.1.2.1.1.1.0

Check whether the '/etc/snmp/snmp/*' files are okay and your snmp-agent
is running ( # ps -ef | grep snmp ). Then make a portscan to port 161
(standard snmp) whether the port is open. If everything is alright you
should get an result.

    Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: New Bee

[Gerhard Ecaroh Froehlich]

Hi Aamir,

"Aamir, Muhammad" wrote:
> 
> Hi Guys,
> 
> Does it matter if I install Solaris 8 or 7 for MRTG. Are there any known
> issues? What should I install to get a better support and less hassle for
> MRTG 2.9.17?

I don't see any problems with 7 or 8. But take care of your compiler and
perl versions. Solaris compiler don't work good for gnu tools like MRTG.
Some versions of Perl don't fit. Make sure you have 'GNU' Environment
;-)) . Check your path's! This is the most common problem on solaris
instead of Linux. Mostly you have to solve dependencies. Take a 'test
machine' for installing GNU Tools and compile it there. After that copy
this to your target but be sure, you have not installed librarys in
other directorys than that one you have been maked.
 
> Thanks.
> 
> Aamir
> email: [EMAIL PROTECTED]


Ecaroh

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: RRD compiling breaks on solaris 2.6 with perl 5.005_03

[Gerhard Ecaroh Froehlich]

Hi,

Peter Kassies wrote:
> 
> Anybody knows what to do about below problem?
> It seems trivial.

It is, it is.

> Peter
> 
> pod2man --release=1.0.33 --center=rrdtool rrdtutorial.es.pod  >
> rrdtutorial.es.1
> /usr/local/bin/pod2man: rrdtutorial.es.pod is missing required section:
> DESCRIPTION
> /usr/bin/nroff -man -Tlp rrdtutorial.es.1 > rrdtutorial.es.txt
> Segmentation Fault - core dumped
> make[1]: *** [rrdtutorial.es.txt] Error 139
> make[1]: Leaving directory `/content/performance/rrdtool-1.0.33/doc'
> make: *** [all-recursive] Error 1

The Spanish Tutorial has some chars in it which does not conform to
standard ASCII Code. Try to open the files with "vi" and you can see the
problem clearly. I copied the english ones over the spanish. Alternately
you can use the other hint with deleting the spanish doc in makefile.

Maybe some programmer on rrdtool might fix the problem in source.
Thanks.

Cheers,

Ecaroh

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: Newbie: rddcgi graph / Apache execution of cgi-scripts

[Gerhard Ecaroh Froehlich]

Hello,

"Jan L. Peterson" wrote:
> 
> The solution is to have your rrd.cgi script in some other directory in
> the doc root.  Add directives to your web server config file to have
> any files ending in .cgi be considered as cgi scripts.  The images
> created will not end in .cgi, so they will be served up as normal image
> files.

I did that so too a few weeks ago. CGI's and graph generation works from
that on. Unfortanetly there are some HTML Files which have names like
"website.cgi.html" and apache tries to execute them. Any ideas?
 
> Good luck.
> 
>     -jan-
> --
>

Thanks in advance,


Gerhard Ecaroh Froehlich

Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: how would you help me? / Links for usefull tools round rrdtool

[Gerhard Ecaroh Froehlich]

Hi zeeshan,

zeeshan ahmed wrote:
> 
> i am new user for rrd tool, i have been using mrtg for last three months now 
> i > want to play with rrd tool, can any one help me in easir way how do i set 
> it 
> up for my routers and switches


I have searched long time for a usefull tool for migrating from MRTG to RRDTool.
Mainly cause i am no programmer at all. At least i have a few links collected
which i can comment after a period of testing.

An very easy way is to still let mrtg do the collecting and add 'use rrdtool' in
mrtg.cfg. This is described in the MRTG/RRDTool Documentation. As data is now
stored in rrdtool and no html and pictures are generated by mrtg anymore you
need to extract it by an cgi. Usefull in this case:

 - 14all.cgi, <http://www.uni-ulm.de/~rbawidam/mrtg-rrd/>
   Usefull, fast to setup but it doesnot scale well in big environents
 - cgi Scripts <http://www.cheshire.demon.co.uk/pub/>
   Takes automaticaly his setup from your existing mrtg.cfg without to much
   configuration. Scales good up to 20 hosts but i had some difficulties 
   to change defaults. Remember, i am not a programmer ;-(
 - cricket, i didnot test this. My first impression was that it does fit 
   network devices better than host performance. But i can be wrong.

For larger environments there are two solutions which use rrdtool without mrtg.
Not all old archives can be migrated:

 - NRG <http://nrg.hep.wisc.edu>
   Seems to work fine on Linux, my Solaris brought some difficulties
 - Remstats <http://remstats.crc.ca>
   Good for very large environments

Without some knowledge round compiling, networking, TCP/IP and SNMP it would be
not simple to set it up. RRDTool is no clicking tool but a scalable flexibel
tool for people who have knowledge in scripting. My experience to commercial
solutions are: Expensive, need lots of work in configuring, does often not fit
special environments. RRDTool is really great! Thanks to Tobi and all
programmers round GNU - good peace of code indeed.
 
> i would be very thank ful to you
> 
> by ebye
> zeeshan ahmed
> Network Engineer
> Worldcall DotCom
> Pakistan
> ph no:09003009414363
> 
> --
> Unsubscribe mailto:[EMAIL PROTECTED]
> Helpmailto:[EMAIL PROTECTED]
> Archive http://www.ee.ethz.ch/~slist/rrd-users
> WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

Sure there are a lot of other good tools round rrdtool and networkmanagment.
Some need lots of experience, some are quiet easy to install. Try it out. If you
have a Unix System most things would be easier, best is Linux at all cause most
of required modules, scripts, compilers are availabel for Linux. I had to set it
up on Solaris - this one is really hard work sometimes. Maybe someone in the
list will reply with additional information. Most of them are on the rrdtool
website as description of 'Frontends' which often mean Backend too to collect
the data.

Cheers, Ecaroh


Gerhard Ecaroh Froehlich, Systemadministrator

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Re: installing RRD-tool problems

[Gerhard Ecaroh Froehlich]

Hello,

"Parchack, Danny" wrote:
> 
> Hi,
> I need the group help installing the Rrdtool on Sun Solaris 8 station:
> During make I received the following message:
> 
> Making all in doc
> pod2man --release=1.0.33 --center=rrdtool rrdtool.pod  > rrdtool.1
> sh: pod2man: not found
> *** Error code 1
> make: Fatal error: Command failed for target `rrdtool.1'
> Current working directory /usr/Danny/rrdtool-1.0.33/doc
> *** Error code 1
> make: Fatal error: Command failed for target `all-recursive'
> 
> thanks
> Danny

You do not have a 'pod2man' in your Path. Mostly this depends on missing GNU
Tools or older perl-versions that ships with standard solaris. Make sure to have
all the nice GNU stuff like GCC, GNUmake and perl >= 5.005 on your system. This
should work. Try 'which pod2man' and 'which make' to find your stuff. If you
install additional GNU software make sure to have them in the right path (echo
$PATH ; setenv PATH /yournewpath:$PATH ; echo $PATH) and _first_ in PATH.

If this doesnot help, mail again.

Cheers, Ecaroh

-- 
Gerhard Ecaroh Froehlich, Systemadministrator 
Real Time Investmentbanking (Team Breitenversorgung)
Telefon: +49 (0) 69 7447 5117
Geno-Rechenzentrum GmbH, 60528 Frankfurt / Main, Hahnstrasse 46

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi

[rrd-users] Frontend for RRDTool, What's your opinion?

[Gerhard Ecaroh Froehlich]

Hello,

excuse my unspecific question. I'm going to migrate from MRTG to RRDTool. I
would be glad to hear some of your opinions about a scalable Frontend (and
collector).

What i have to monitor is:
 - over 100 Sun Solaris Server
 - 100's of custom made processes and binaries giving custom output
 - doing real-time dataprocessing distributed via Satellite

Of course, there is a need having a scalable frontend and data collector
(SNMP/custom scripts) without reeventing the wheel. And also code which will be
a startingpoint with clear and functionable code on which i can rely and use to
extend it for my needs. I spent some time in evaluating 14all, cricket,
RRGrapher and similar tools. But time is progressing cause every test results in
troubleshooting for compiling, installing different modules and configuring till
tests succeed. So, what experiences did you made with RRDTool Frontends in large
environments? Is there your prefered solutions which fit such needs?

Any comments and sugestions would be highly appreciated. Feel free to mail
directly to my address if you feel that my wish is not interesting for the list.
Thank you all.


Cheers, Ecaroh


-- 
Gerhard Ecaroh Froehlich
Systems-Administrator 
Real Time Investmentbanking
Geno-Rechenzentrum GmbH

--
Unsubscribe mailto:[EMAIL PROTECTED]
Helpmailto:[EMAIL PROTECTED]
Archive http://www.ee.ethz.ch/~slist/rrd-users
WebAdminhttp://www.ee.ethz.ch/~slist/lsg2.cgi