Koozali.org formerly Contribs.org

Announcement: Howto monitor your sme7 general performance us

Announcement: Howto monitor your sme7 general performance us
« on: October 24, 2005, 11:04:35 PM »
Hi,
I have made this "Howto monitor your sme7 general performance using with RRDtool and collectd" available at:
http://www.hanscees.com/sme7/collectd4sme7howto.html

Here is what you get: take a look!
http://www.hanscees.com/sme7/collection.cgi.html

What this howto will and will not produce:

    * 1. This howto works in that it produces RRD-tool graphics of performance relevant data like cpu (load), memory usage, IO performance and interface statistics (traffic flow).
    * 2. This howto is not yet finished/polished toward sme. Anyone is called upon to improve it. I will be willing to help if needed.
    * 3. This howto should also generally work for centhos4 and redhat 4 enterprise, since sme7 is based upon that.

Special thanks to Florian Forster (of collectd ) for helping me out with some initial troubleshooting. And of course to the volunteers of sme7 (www.contribs.org) for making a great platform.





Hope it is of use to some. I am sure it can be improved. Please let me in on your good ideas.

greetings
Hans-Cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #1 on: October 27, 2005, 12:25:33 AM »
Hi
if anyone uses this howto and has an opinion about it I sure would like to know.

cheers

Hans-Cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #2 on: October 27, 2005, 07:14:00 PM »
Hi Hans,

Just found the HowTo and followed it on my SME7 testbox.
Nice work ... :-D :-D :-D
I do get the results you show on your site with little effort a fe minutes later.

All in all this sure seems to have to potential to replace/follow the defunct sysmon in SME 7.

I have had no time to dig deeper so 2 questions:
- Did you test the lm_sensors part? I could not (yet) get this to display a correct .rrd graph for temps and fan speed ("sensors" itself does work correctly).
- Since I support a lot of Raid5 setups I find myself using hddtemp very often. Surprisingly Version 3.2.2 of collectd does support this. Do you have an updated rpm already .-)

Again: Nice work.

Regards
Reinhold
............

Announcement: Howto monitor your sme7 general performance us
« Reply #3 on: October 28, 2005, 12:03:18 AM »
Quote from: "Reinhold"

I have had no time to dig deeper so 2 questions:
- Did you test the lm_sensors part? I could not (yet) get this to display a correct .rrd graph for temps and fan speed ("sensors" itself does work correctly).
- Since I support a lot of Raid5 setups I find myself using hddtemp very often. Surprisingly Version 3.2.2 of collectd does support this. Do you have an updated rpm already .-)

Again: Nice work.

Regards
Reinhold


Thanks!
I have not done anything to the package itself. I have not tested sensors.
I have produced new rpm's, but did not test them at all. If you want please test them.


http://www.hanscees.com/sme7/collectd-3.2.0-1.i386.rpm
http://www.hanscees.com/sme7/collectd-sensors-3.2.0-1.i386.rpm
http://www.hanscees.com/sme7/collectd-hddtemp-3.2.0-1.i386.rpm


greetings

Hans-cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #4 on: October 28, 2005, 06:38:39 PM »
Hi Hans,

Had little time so I hope for "a rainy weekend"  ;-)

I did manage a short look into the sensors.c file however and indeed it says:
 /* Only temperature for now.. */
and it seemingly tries the wrong sensor values for this PC.
I will look into this over the weekend ...

Thanks for the updated rpms. I'll grab them and report as to the functionality of hddtemp ... asap.
:pint:

Regards
Reinhold
............

Announcement: Howto monitor your sme7 general performance us
« Reply #5 on: October 28, 2005, 07:53:02 PM »
Hi Hans,

Short feedback: No hddtemp data :-(

longer version:
Code: [Select]
rpm -Uvh collectd-3.2.0-1.i386.rpm
Error: Failed dependencies:
        perl(XML::Dumper) is needed by collectd-3.2.0-1.i386
        collectd = 3.0.0 is needed by (installed) collectd-sensors-3.0.0-2.i386


... so there are some spec ? problems still.

I decided to install the "official Centos" XML-Dumper

Code: [Select]
wget ftp://ftp.nluug.nl/pub/os/Linux/distr/CentOS/4.2/os/alpha/CentOS/RPMS/perl-XML-Dumper-0.71-2.noarch.rpm
rpm -Uvh perl-XML-Dumper-0.71-2.noarch.rpm


... and (see above) since collectd (falsely ?!) needs still it's predecessor 3.0.0 I went nodeps like you did...

Code: [Select]
rpm -Uvh --nodeps collectd-3.2.0-1.i386.rpm
Preparing...                ########################################### [100%]
   1:collectd               ########################################### [100%]

rpm -Uvh collectd-hddtemp-3.2.0-1.i386.rpm
Preparing...                ########################################### [100%]
   1:collectd-hddtemp       ########################################### [100%]

 rpm -Uvh collectd-sensors-3.2.0-1.i386.rpm
Preparing...                ########################################### [100%]
   1:collectd-sensors       ########################################### [100%]


the new cgi is obviously needed so (with the directories like in your howto)

Code: [Select]
cp /usr/share/doc/collectd-3.2.0/contrib/museum/collection.cgi /home/e-smith/files/ibays/stats/cgi-bin/collection.cgi

... same data as before... collectd does not generate hddtemp-data
...but hddtemp code was installed

Code: [Select]
rpm -ql collectd-hddtemp
/usr/lib/collectd/hddtemp.so
/usr/lib/collectd/hddtemp.so.0
/usr/lib/collectd/hddtemp.so.0.0.0


Note that the "-p www.contribs.org" does give a .rrd but not graph :-(

I am in a bit of a rush now...
just wanted to give the test of the hddtemp rpms (thanks for the work)
...more over the weekend (hopefully)

Let's polish this stuff ;-)

Regards
Reinhold

P.S.: For one reason or another I do have sensor data for TEMP and FANSPEED as .rrd now (3 graphs which is the correct number ... but the data is still invalid)
............

Announcement: Howto monitor your sme7 general performance us
« Reply #6 on: October 28, 2005, 08:33:47 PM »
Quote from: "Reinhold"
Hi Hans,

Short feedback: No hddtemp data :-(

longer version:
Code: [Select]
rpm -Uvh collectd-3.2.0-1.i386.rpm
Error: Failed dependencies:
        perl(XML::Dumper) is needed by collectd-3.2.0-1.i386
        collectd = 3.0.0 is needed by (installed) collectd-sensors-3.0.0-2.i386


... so there are some spec ? problems still.



But you should update collectd-sensors-3.0.0-2.i386 with the newer version of sensors as well I think.?

Quote


I decided to install the "official Centos" XML-Dumper

Code: [Select]
wget ftp://ftp.nluug.nl/pub/os/Linux/distr/CentOS/4.2/os/alpha/CentOS/RPMS/perl-XML-Dumper-0.71-2.noarch.rpm
rpm -Uvh perl-XML-Dumper-0.71-2.noarch.rpm


... and (see above) since collectd (falsely ?!) needs still it's predecessor 3.0.0 I went nodeps like you did...

Code: [Select]
rpm -Uvh --nodeps collectd-3.2.0-1.i386.rpm
Preparing...                ########################################### [100%]
   1:collectd               ########################################### [100%]

rpm -Uvh collectd-hddtemp-3.2.0-1.i386.rpm
Preparing...                ########################################### [100%]
   1:collectd-hddtemp       ########################################### [100%]

 rpm -Uvh collectd-sensors-3.2.0-1.i386.rpm
Preparing...                ########################################### [100%]
   1:collectd-sensors       ########################################### [100%]



the new cgi is obviously needed so (with the directories like in your howto)



The collection.cgi is fine for a start, but the maintainer shows little interest in beefing it up at this moment I am afraid.

So I am looking for a good perl hacker to change that script. I have been staring at it, but do not understand the code I am afraid.

Quote




Code: [Select]
cp /usr/share/doc/collectd-3.2.0/contrib/museum/collection.cgi /home/e-smith/files/ibays/stats/cgi-bin/collection.cgi

... same data as before... collectd does not generate hddtemp-data
...but hddtemp code was installed

Code: [Select]
rpm -ql collectd-hddtemp
/usr/lib/collectd/hddtemp.so
/usr/lib/collectd/hddtemp.so.0
/usr/lib/collectd/hddtemp.so.0.0.0


Note that the "-p www.contribs.org" does give a .rrd but not graph :-(


another collection.cgi issue.
Quote



I am in a bit of a rush now...
just wanted to give the test of the hddtemp rpms (thanks for the work)
...more over the weekend (hopefully)

Let's polish this stuff ;-)

Regards
Reinhold

P.S.: For one reason or another I do have sensor data for TEMP and FANSPEED as .rrd now (3 graphs which is the correct number ... but the data is still invalid)


Perhaps we should just drop collection.cgi and make a new script, but that seems such a waste.

greetings

Hans-Cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #7 on: October 29, 2005, 06:47:20 PM »
Hans,

I did a bit of fumbling ...

- A new lm_sensors setup rpm + search a la wellsi.com ... and whatshallIsay ...works now !
I get Case TEMP & CPU TEMP (temp1 and temp2), and CPU FAN Data.
And -(un)happy surprise- I found out that the test-servers cpu-fan has intermittend problems ...
see below... IMO that alone makes a tool like this invaluable.
Note: Fan-scale is Celsius should read RPM -cosmetic-

- A bit of "c"ing into the code and ...Hddtemp does need the daemon mode i.e. you need to start it before "connected"-ing:
Code: [Select]
hddtemp -d /dev/hda
...so it connects now but data is still wrong for the test HD  :cry:

- "Processes" currently gives no .rrd output whatsoever :-?

- Ping logging as in :
Code: [Select]
collected -p www.contribs.org
does work too. See below how "beautifully bad" it pinged last night (while europeans should be sleeping) :-D

- Note: This is an SME7 server-only so no bandwith (yet) on eth1...

AND NOW IT DOES GET "COLORFUL" :roll:
- Although my perl is rather "pedestrian", tonight I cut&paste-ed my way to a output.cgi that does please me enough to continue...

Now here are the graphs in all their "splendor"
(Folks please note ... this is not on contribs so it will only strain any viewer's_&_photobucket's not Ruffdog's bandwitdth)  8-)














Right now the display above is rather special for my test SME7...

(I only display the .rrd files that collectd places in /var/lib/collectd/ with a cut&paste version of collection.cgi)
... but...
with some fiddling I feel able to make this universal enough for display on a server-manager page.

Regards
Reinhold

P.S.: Had a short glimpse on your updated website ... nice going!
............

Announcement: Howto monitor your sme7 general performance us
« Reply #8 on: October 29, 2005, 11:08:50 PM »
Quote from: "Reinhold"
Hans,

I did a bit of fumbling ...

- A new lm_sensors setup rpm + search a la wellsi.com ... and whatshallIsay ...works now !


a la welsii.com??

Quote



I get Case TEMP & CPU TEMP (temp1 and temp2), and CPU FAN Data.
And -(un)happy surprise- I found out that the test-servers cpu-fan has intermittend problems ...
see below... IMO that alone makes a tool like this invaluable.
Note: Fan-scale is Celsius should read RPM -cosmetic-

- A bit of "c"ing into the code and ...Hddtemp does need the daemon mode i.e. you need to start it before "connected"-ing:
Code: [Select]
hddtemp -d /dev/hda



Could you state that a little more explicit? Do you need to give the hddtmp command before running collectd?
Doesn't collectd start the hddtmp up?

Quote



...so it connects now but data is still wrong for the test HD  :cry:

- "Processes" currently gives no .rrd output whatsoever :-?

- Ping logging as in :
Code: [Select]
collected -p www.contribs.org
does work too. See below how "beautifully bad" it pinged last night (while europeans should be sleeping) :-D

- Note: This is an SME7 server-only so no bandwith (yet) on eth1...

AND NOW IT DOES GET "COLORFUL" :roll:
- Although my perl is rather "pedestrian", tonight I cut&paste-ed my way to a output.cgi that does please me enough to continue...


could you post it or somewhere? Or mail it to me? My perl is probably worse than yours.
I asked Floria Foster from collectd to "upgrade " the scirpt. See the collectd mailing list at
http://mailman.verplant.org/listinfo/collectd

I would like a script that gives you the choice of a time frame, so you can see both day/week/month data. That is good to spot trends. I know howto do that with  RRD,
see
http://www.macsat.com/rrdtool.php
and the script that leads to.



Quote

(I only display the .rrd files that collectd places in /var/lib/collectd/ with a cut&paste version of collection.cgi)
... but...
with some fiddling I feel able to make this universal enough for display on a server-manager page.


that would be great indeed. I feel comfertable to make a howto, but hacking server-manager is beyond what I can do now. I just don't understand how it works yet.

Quote



Regards
Reinhold

P.S.: Had a short glimpse on your updated website ... nice going!


Thanks, it was a bit late when I finished it:-)

Hans-Cees
nl.linkedin.com/in/hanscees/

matsk

Announcement: Howto monitor your sme7 general performance us
« Reply #9 on: October 30, 2005, 12:47:04 AM »
To simplify the installation of cpan modules you could use
Code: [Select]
# install perl-xml-dumper
cpan install XML::Dumper


instead of the code :
Code: [Select]
wget http://ftp.debian.org/debian/pool/main/libx/libxml-dumper-perl/libxml-dumper-perl_0.79.orig.tar.gz

# install perl-xml-dumper
tar -zxvf libxml-dumper-perl_0.79.orig.tar.gz
cd XML-Dumper-0.79
perl Makefile.PL
make
make test
make install


The only thing that could be confusing is that the first time cpan is executed it has to be configured.
The advantage with cpan is that it takes care of dependencies in the same matter that yum does and will keep your cpan modules updated and at the latest version.

/Mats
PS. And hanscees: could you update your webpage "script" with the latest rpm's that you have buildt ?

matsk

Announcement: Howto monitor your sme7 general performance us
« Reply #10 on: October 30, 2005, 01:06:52 AM »
If you would like to have a "short URL" you could create a index.html
Code: [Select]
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<html>
<head>
<title>SME statistics</title>
<meta http-equiv="REFRESH" content="0;url=http://the ip or hostname/stats/cgi-bin/collection.cgi"></HEAD>
<BODY>

</BODY>
</HTML>


This will redirect you to the cgi script and you dont have to remember it, just surf to "http://the ip or hostname/stats/"

Announcement: Howto monitor your sme7 general performance us
« Reply #11 on: October 30, 2005, 10:37:15 AM »
Hans,

The damned CPU-Fan really died on me... so I lost access to my test SME7 ... at least for the weekend :cry:

As to your other questions:

- Ian Wells, very active contribs member (Thanks Ian!), has done a lot on lm_sensors see here
This will tell you how to find and activate the sensors in your PC as used by collected.
Note: SME7 does include a working version of lm_sensors - a new rpm install should not be needed ...


- hddtemp needs to be installed separately.
Best get it from the DAG Centos repo - or a mirror since it seems down/busy a lot recently. See here
After installation you can start hddtemp in two different modes.
First mode gives the temp data directly, second mode goes into daemon state (background)
and you can access the data via a port on the pc locally, try (ssh'd into SME7)
Code: [Select]
hddtemp -d /dev/hd[abcd]
telnet localhost 7634

man hddtemp will tell you in detail what this means and a look into hddtemp.c from collectd shows  what it should do :-x  


Last not least:
sysmon ...on SME 6.0 does give more and better organized data in the fashion you (too) would like to see, namely day/month/year values.  
Have a look. I do have the rpm source available locally or here (?)and this is not on the SME7 test server :-)  
(Note: original was from eneo/ShadLord so the above may be incorrectly googled.)
- Maybe best for us to look into the code there before cutting&pasting more
Perl... which I (still) sort of dislike for being so ... "unforgiving" :-(

Regards
Reinhold

matsk:
Perl pro? Please have a look into sysmon src as above!
(lazy like I tend to be) I already did forward my own smestats/index.html (like you proposed here) but
IMHO collectd primarily has to go into a server-manager-link to be accessible... and it has to refresh there too ;-)
............

Announcement: Howto monitor your sme7 general performance us
« Reply #12 on: October 30, 2005, 08:12:04 PM »
Thanks for the FAQ Hans. It worked fine for me on 7.0b4.

Couple of questions...

I added the new 3.2 collectd rpm's and hddtemp as you posted above, but no new graphs appeared on my ibay. Do I have to manually add hddtemp into the stats somehow? I do have a set of 3 hddtemp.so files in /usr/lib/collectd and hddtemp is running as a daemon.

telnet localhost 7634
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
|/dev/hda|Maxtor 6B200P0|29|C|
Connection closed by foreign host.


Running 'cpan install XML::Dumper' as matsk suggested caused a huge amount of errors. I don't know why but it seems to be a gcc path issue?

[root@host cpustats]# cpan install XML::Dumper
CPAN: Storable loaded ok
Going to read /root/.cpan/Metadata
  Database was generated on Sun, 30 Oct 2005 08:13:39 GMT
Running install for module install
Running make for J/JU/JUNOS/junoscript-perl-6.4I0.tar.gz
CPAN: Digest::MD5 loaded ok
CPAN: Compress::Zlib loaded ok
Checksum for /root/.cpan/sources/authors/id/J/JU/JUNOS/junoscript-perl-6.4I0.tar.gz ok
Scanning cache /root/.cpan/build for sizes
junoscript-perl-6.4I0/
[snip]
Removing previously used /root/.cpan/build/junoscript-perl-6.4I0

  CPAN.pm: Going to build J/JU/JUNOS/junoscript-perl-6.4I0.tar.gz

Checking if your kit is complete...
Looks good
Warning: prerequisite IO::Tty 0 not found.
Writing Makefile for junoscript-perl
---- Unsatisfied dependencies detected during [J/JU/JUNOS/junoscript-perl-6.4I0.tar.gz] -----
    IO::Tty
Shall I follow them and prepend them to the queue
of modules we are processing right now? [yes]
Running make test
  Delayed until after prerequisites
Running make install
  Delayed until after prerequisites
Running install for module IO::Tty
Running make for R/RG/RGIERSIG/IO-Tty-1.02.tar.gz
Checksum for /root/.cpan/sources/authors/id/R/RG/RGIERSIG/IO-Tty-1.02.tar.gz ok
IO-Tty-1.02/
IO-Tty-1.02/Pty.pm
IO-Tty-1.02/Makefile.PL
IO-Tty-1.02/ChangeLog
IO-Tty-1.02/test.pl
IO-Tty-1.02/Tty.xs
IO-Tty-1.02/Tty.pm
IO-Tty-1.02/README
IO-Tty-1.02/try
IO-Tty-1.02/MANIFEST
Removing previously used /root/.cpan/build/IO-Tty-1.02

  CPAN.pm: Going to build R/RG/RGIERSIG/IO-Tty-1.02.tar.gz

Now let's see what we can find out about your system
(logfiles of failing tests are available in the conf/ dir)...

ERROR: cannot run the configured compiler 'gcc'
(see conf/compilerok.log).  Please fix this by adjusting
perls Config.pm, creating a symlink to the right location
or maybe by just adding it to your PATH.  You could also build
perl yourself.

Running make test
  Make had some problems, maybe interrupted? Won't test
Running make install
  Make had some problems, maybe interrupted? Won't install
Running make for J/JU/JUNOS/junoscript-perl-6.4I0.tar.gz
  Is already unwrapped into directory /root/.cpan/build/junoscript-perl-6.4I0

  CPAN.pm: Going to build J/JU/JUNOS/junoscript-perl-6.4I0.tar.gz

cp lib/JUNOS/Access/clear_text.pm blib/lib/JUNOS/Access/clear_text.pm
cp lib/JUNOS/Access/stubs.pm blib/lib/JUNOS/Access/stubs.pm
cp lib/JUNOS/Methods.pm blib/lib/JUNOS/Methods.pm
cp lib/JUNOS/Access.pm blib/lib/JUNOS/Access.pm
cp lib/JUNOS/Response.pm blib/lib/JUNOS/Response.pm
cp lib/JUNOS/Access/ssh.pm blib/lib/JUNOS/Access/ssh.pm
cp lib/JUNOS/Device.pm blib/lib/JUNOS/Device.pm
cp lib/JUNOS/6.4I0/jkernel_methods.pl blib/lib/JUNOS/6.4I0/jkernel_methods.pl
cp lib/JUNOS/version.pm blib/lib/JUNOS/version.pm
cp lib/JUNOS/6.4I0/jcrypto_methods.pl blib/lib/JUNOS/6.4I0/jcrypto_methods.pl
cp lib/JUNOS/Access/xnm.pm blib/lib/JUNOS/Access/xnm.pm
cp lib/JUNOS/Trace.pm blib/lib/JUNOS/Trace.pm
cp lib/JUNOS/6.4I0/jroute_methods.pl blib/lib/JUNOS/6.4I0/jroute_methods.pl
cp lib/JUNOS/Access/rsh.pm blib/lib/JUNOS/Access/rsh.pm
cp lib/JUNOS/Access/ssl.pm blib/lib/JUNOS/Access/ssl.pm
cp lib/JUNOS/Access/telnet.pm blib/lib/JUNOS/Access/telnet.pm
cp lib/JUNOS/DOM/Parser.pm blib/lib/JUNOS/DOM/Parser.pm
Manifying blib/man3/JUNOS::Access::clear_text.3pm
Manifying blib/man3/JUNOS::Access::stubs.3pm
Manifying blib/man3/JUNOS::Methods.3pm
Manifying blib/man3/JUNOS::Access.3pm
Manifying blib/man3/JUNOS::Access::ssh.3pm
Manifying blib/man3/JUNOS::Response.3pm
Manifying blib/man3/JUNOS::Access::rsh.3pm
Manifying blib/man3/JUNOS::Access::ssl.3pm
Manifying blib/man3/JUNOS::Device.3pm
Manifying blib/man3/JUNOS::Access::telnet.3pm
Manifying blib/man3/JUNOS::Trace.3pm
Manifying blib/man3/JUNOS::Access::xnm.3pm
Manifying blib/man3/JUNOS::DOM::Parser.3pm
  /usr/bin/make  -- OK
Running make test
PERL_DL_NONLAZY=1 /usr/bin/perl "-MExtUtils::Command::MM" "-e" "test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/base-1....Can't locate IO/Pty.pm in @INC (@INC contains: /root/.cpan/build/junoscript-perl-6.4I0/blib/lib/5.8.5/i386-linux-thread-multi
[snip]

/usr/lib/perl5/vendor_perl/5.8.0 /usr/lib/perl5/vendor_perl .) at /root/.cpan/build/junoscript-perl-6.4I0/blib/lib/JUNOS/Access.pm line 41.
BEGIN failed--compilation aborted at /root/.cpan/build/junoscript-perl-6.4I0/blib/lib/JUNOS/Access.pm line 41.
Compilation failed in require at /root/.cpan/build/junoscript-perl-6.4I0/blib/lib/JUNOS/Device.pm line 45.
BEGIN failed--compilation aborted at /root/.cpan/build/junoscript-perl-6.4I0/blib/lib/JUNOS/Device.pm line 45.
Compilation failed in require at t/base-2.t line 12.
BEGIN failed--compilation aborted at t/base-2.t line 12.
t/base-2....dubious                                                          
        Test returned status 2 (wstat 512, 0x200)
FAILED--2 test scripts could be run, alas--no output ever seen
make: *** [test_dynamic] Error 2
  /usr/bin/make test -- NOT OK
Running make install
  make test had returned bad status, won't install without force
XML::Dumper is up to date.

Announcement: Howto monitor your sme7 general performance us
« Reply #13 on: October 30, 2005, 09:12:38 PM »
Quote from: "matsk"
To simplify the installation of cpan modules you could use
Code: [Select]
# install perl-xml-dumper
cpan install XML::Dumper


instead of the code :
Code: [Select]
wget http://ftp.debian.org/debian/pool/main/libx/libxml-dumper-perl/libxml-dumper-perl_0.79.orig.tar.gz

# install perl-xml-dumper
tar -zxvf libxml-dumper-perl_0.79.orig.tar.gz
cd XML-Dumper-0.79
perl Makefile.PL
make
make test
make install


The only thing that could be confusing is that the first time cpan is executed it has to be configured.
The advantage with cpan is that it takes care of dependencies in the same matter that yum does and will keep your cpan modules updated and at the latest version.

/Mats
PS. And hanscees: could you update your webpage "script" with the latest rpm's that you have buildt ?


That might work very nicely. But it could stab you in the back badly too. cpan will get you a new perl version if some dependency asks it too. I would not recommend this in a production environment at all. I would choose the rpm, in which case you do not need to install I think. Just do rpm -ihv ... I think

I am updating my howto, but not with everything. I want to keep it stable. If we have something better that is tested and works I will gladly post it there.

Hans-Cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #14 on: October 30, 2005, 10:14:35 PM »
Quote from: "Reinhold"
Hans,

- Ian Wells, very active contribs member (Thanks Ian!), has done a lot on lm_sensors see here
This will tell you how to find and activate the sensors in your PC as used by collected.
Note: SME7 does include a working version of lm_sensors - a new rpm install should not be needed ...


I have followed this lm_sensors stuff.
In sme7 lm_sensors is installed. But it does not seem to have all kinds of via modules.
Can anynbody else check if their sme7 has lm_sensor modules? I cannot find vt1211 nor vt8235. I have a via -board (via pd) and should copy a sensor module. But I do not know where to copy it from .

This site:
http://seanm.ca/eden/lmsensors.html

says:
Copy the vt8231 file to /lib/modules/uname -r/kernel/drivers/sensors.

But I do not know where to get that module. It should be in lm_sensors already if I understand it right, since it is supported bu lm_sensors.

Hans-Cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #15 on: October 30, 2005, 11:21:27 PM »
Quote from: "Reinhold"
Hans,

- Ian Wells, very active contribs member (Thanks Ian!), has done a lot on lm_sensors see here
This will tell you how to find and activate the sensors in your PC as used by collected.
Note: SME7 does include a working version of lm_sensors - a new rpm install should not be needed ...


I have followed this lm_sensors stuff.
In sme7 lm_sensors is installed. But it does not seem to have all kinds of via modules.
Can anynbody else check if their sme7 has lm_sensor modules? I cannot find vt1211 nor vt8235. I have a via -board (via pd) and should copy a sensor module. But I do not know where to copy it from .

This site:
http://seanm.ca/eden/lmsensors.html

says:
Copy the vt8231 file to /lib/modules/uname -r/kernel/drivers/sensors.

But I do not know where to get that module. It should be in lm_sensors already if I understand it right, since it is supported bu lm_sensors.

I got hddtemp to work just fine for asking it stuff via  telnet. But there is no data in the graphic output of collection.cgi. Do you have a line?

The processes graphic is not there. That has to do with that it has no defintion in the cgi script.



Hans-Cees[/quote]
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #16 on: October 31, 2005, 07:50:05 AM »
Hans

Are you seeing hddtemp on your install? I am not - what am I doing wrong?

Morph

Announcement: Howto monitor your sme7 general performance us
« Reply #17 on: October 31, 2005, 09:28:15 PM »
Quote from: "morpheusx"
Hans

Are you seeing hddtemp on your install? I am not - what am I doing wrong?

Morph


you need to install it and configure/test it. See the bottom of the howto. You need to use the latest collectd rpm's too I think.

Hans-Cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #18 on: November 01, 2005, 12:08:04 AM »
Quote from: "hanscees"
Quote from: "morpheusx"
Hans

Are you seeing hddtemp on your install? I am not - what am I doing wrong?

Morph


you need to install it and configure/test it. See the bottom of the howto. You need to use the latest collectd rpm's too I think.

Hans-Cees


I do have all the latest rpms installed - but collectd does not show hddtemp in the CGI output on the web page.

[root@host ~]# rpm -q collectd-hddtemp
collectd-hddtemp-3.2.0-1
[root@host ~]# rpm -q collectd
collectd-3.2.0-1
[root@host ~]# rpm -q collectd-sensors
collectd-sensors-3.2.0-1
[root@host ~]# rpm -q hddtemp
hddtemp-0.3-0.beta12.2.2.el4.rf
[root@host ~]# telnet localhost 7634
Trying 127.0.0.1...
Connected to localhost.
Escape character is '^]'.
|/dev/hda|Maxtor 6B200P0|29|C||/dev/hdb|Maxtor 6L200R0|29|C|Connection closed by foreign host.

There are hddtemp files in the collectd directory but for some reason they aren't turned into a web output

ls /usr/lib/collectd

cpu.so        disk.so.0      hddtemp.so.0.0.0  memory.so        ping.so.0       processes.so.0.0.0  swap.so        traffic.so.0
cpu.so.0      disk.so.0.0.0  load.so           memory.so.0      ping.so.0.0.0   sensors.so          swap.so.0      traffic.so.0.0.0
cpu.so.0.0.0  hddtemp.so     load.so.0         memory.so.0.0.0  processes.so    sensors.so.0        swap.so.0.0.0
disk.so       hddtemp.so.0   load.so.0.0.0     ping.so          processes.so.0  sensors.so.0.0.0    traffic.so

Does anyone have hddtemp displaying in the collectd web output?

Announcement: Howto monitor your sme7 general performance us
« Reply #19 on: November 01, 2005, 02:28:56 PM »
Hans,
Try
Code: [Select]
wget ftp://ftp.nluug.nl/pub/os/Linux/distr/CentOS/4.2/os/i386/CentOS/RPMS/lm_sensors-devel-2.8.7-2.i386.rpm
rpm -Uvh lm_sensors-devel-2.8*

(this should include the via sensors you listed). Then follow "wellsi" ...

morpheux,
Yes I did have hddtemp  rrd output - and it will be displayed with the new (!) collection.cgi.
Reread what I said above ... you do need to copy the new collection.cgi coming with collectd 3.2.x also.
BUT: the data displayed is bad ... some problems with  hddtemp.c (?!)
BTW: The data ready for display is located in /var/lib/collectd/ !

Regards
Reinhold
...waiting for a new FAN+CPU for his SME7 test box :-(
............

Announcement: Howto monitor your sme7 general performance us
« Reply #20 on: November 01, 2005, 06:08:47 PM »
Quote from: "Reinhold"
Hans,

Reread what I said above ... you do need to copy the new collection.cgi coming with collectd 3.2.x also.
BUT: the data displayed is bad ... some problems with  hddtemp.c (?!)


Got it - thanks!

cp /usr/share/doc/collectd-3.2.0/contrib/museum/collection.cgi /home/e-smith/files/ibays/cpustats/cgi-bin/collection.cgi

and now it works! The drive temps look fine on my system - they are in Celsius but appear to be accurate.

Announcement: Howto monitor your sme7 general performance us
« Reply #21 on: November 01, 2005, 10:27:20 PM »
Quote from: "Reinhold"
Hans,
Try
Code: [Select]
wget ftp://ftp.nluug.nl/pub/os/Linux/distr/CentOS/4.2/os/i386/CentOS/RPMS/lm_sensors-devel-2.8.7-2.i386.rpm
rpm -Uvh lm_sensors-devel-2.8*

(this should include the via sensors you listed). Then follow "wellsi" ...


I installed that rpm but tough luck I am afraid:
[root@mail ~]# find / -name vt1211.ko
[root@mail ~]# modprobe vt1211
FATAL: Module vt1211 not found.

On another point:
I did ask collectd developer to adjust the processes graph. The problem is I think that the specs for that graph are not in collection.cgi.

Hans-Cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #22 on: November 02, 2005, 10:43:09 AM »
Hans,

Oops ... Seems like a problem of the centos/SME7 kernel ...
(since that lm_sensors version DOES support the chip)

Quote from: "- markus.dolic.com -"
...of course the latest » lm_sensors version didn’t just work out-of-the-box: sensors-detect found a VT1211 Super IO Sensors chip which needs the » vt1211 module (among some i2c things). as vt1211 is not part of the kernel yet, it needs to be built as an external module. There’s a kernel patch file on the aforementioned vt1211 page, but i copied vt1211.c and changed the corresponding files manually because the patch assumed the kernel source to be in some other directory than /usr/src/linux (just 2 files anyway).

After setting CONFIG_SENSORS_VT1211=m in .config and running make -C /usr/src/linux M=pwd modules followed by modules_install in the drivers/i2c/chips directory, loading the module (insmod /path/to/vt1211.ko) surprisingly worked without any problems and the sensors program showed a bunch of values (kernel version 2.6.11.6, by the way).



See:  http://hem.bredband.net/ekmlar/vt1211.html for a detailed fix.

Regards
Reinhold
............

persisto

sme6admin
« Reply #23 on: November 02, 2005, 08:02:05 PM »
Hello,

I find your subject intersting, and am curiouis to know the differences between your solution and the one proposed in:
http://http://firewall-services.com/sme6admin-howto-fr.html
It seemed to work well in v6.0 and claims to 7.0 compatible since recent.

Thanks

Announcement: Howto monitor your sme7 general performance us
« Reply #24 on: November 02, 2005, 10:00:54 PM »
persisto,

4me two reasons:

i)  Qu'est-ce qu'il ne fait pas ?:  marche  8-)
iow Didn't work for me when I checked in 7beta5...
Code: [Select]
Can't call method "value" on an undefined value at /etc/e-smith/web/panels/manager/cgi-bin/sme6admin line 182.

ii) collectd is a very fast compiled c programm - 10 measurements/second in all the graph's as shown above.
It still needs a bunch of polishing but it does measure more already .-) and it's just a couple of weeks old!

But I do admit that I did not use sme6admin in 6.x (used sysmon) and I might be quite willing to give it another try
... latest when it's called SME7admin :-D


Regards
Reinhold
............

Announcement: Howto monitor your sme7 general performance us
« Reply #25 on: November 07, 2005, 09:43:16 PM »
Hi,

I have found howto hack collection.cgi into giving you hour/day/week/month graphs:

Beste copy the script to collectionall.cgi.
Find lines:

            if (length ($inst))
            {
                src="$MySelf$RelDir/$type/$inst/day" /><br />\n);

and add some lines so it becomes:

            if (length ($inst))
            {
                print qq(<img src="$MySelf$RelDir/$type/$inst/3600" /><br />\n);
                print qq(<img src="$MySelf$RelDir/$type/$inst/day" /><br />\n);
                print qq(<img src="$MySelf$RelDir/$type/$inst/week" /><br />\n);
                print qq(<img src="$MySelf$RelDir/$type/$inst/month" /><br />\n);


thats it!


Hans-Cees
nl.linkedin.com/in/hanscees/

nitrouze

Announcement: Howto monitor your sme7 general performance us
« Reply #26 on: November 11, 2005, 12:47:15 PM »
Hi,
hanscees, I was wondering if you find a way regarding the vt1211 modules as I'm also running sme on an epia ITX board ?

Keep on the good job

Announcement: Howto monitor your sme7 general performance us
« Reply #27 on: November 11, 2005, 06:56:31 PM »
Quote from: "nitrouze"
Hi,
hanscees, I was wondering if you find a way regarding the vt1211 modules as I'm also running sme on an epia ITX board ?

Keep on the good job


No. I am afraid not. There are working versions it seems, but I have not been able to track them.

Here is some info I found where the module is named:
http://www.epiawiki.org/wiki/tiki-index.php?page=EpiaInstallingGentoo

I was not able to get the code we need from it though.

Perhaps we should ask redhat bugzilla if they can incorporate it. But we would need a place to point at where working code/patches are.

Hans-Cees
nl.linkedin.com/in/hanscees/

Announcement: Howto monitor your sme7 general performance us
« Reply #28 on: December 21, 2005, 06:42:15 AM »
I have follwed this thread very closely and implemented the howto exactly but Iam unable to get the stats to display.

It appears data is being collected but this is all I can get when I try to display the data.

Internal Server Error
The server encountered an internal error or misconfiguration and was unable to complete your request.

Please contact the server administrator, admin@

I am really not sure where to go from here. Any suggestions appreciated

Thanks
Peter
...

Announcement: Howto monitor your sme7 general performance us
« Reply #29 on: December 21, 2005, 06:46:47 AM »
I guess it might help if I say I am running SME7.0B8
...

Announcement: Howto monitor your sme7 general performance us
« Reply #30 on: December 21, 2005, 12:55:03 PM »
After pouring over this for a number of hours and many searches, I changed the permissions on the collection & museum scripts from 660 to 775 and voila - it burst into life and looking like what I was expecting !!
...

Announcement: Howto monitor your sme7 general performance us
« Reply #31 on: December 22, 2005, 12:05:20 AM »
When you get a server-error please have a look at the apache error log in /var/log/httpd I think. Some stuff took me a long time too to find out:-(

Glad it works now.

Hans-Cees
nl.linkedin.com/in/hanscees/

eborgard

vt1211 and SME v7
« Reply #32 on: November 11, 2006, 02:59:26 PM »
I am using a M10000 EPIA board. I also have the problem of the missing vt1211 driver. Has anyone found a way to patch and compile the vt1211.c file and make it work with SME v7?