perfSONAR User Q&A and Other Discussion

[Andrew Lake <>]

Hi,

Looks like you have multiple things going on. See inline comments below:

On Apr 24, 2014, at 11:15 PM, Phil Reese 
<>
 wrote:

> Hi Andy,
> 
> That pointer did clear up the spinning wheel problem but I'm still stuck 
> with three systems showing appropriate dashboards, but no results, all 
> orange.
> 
> Here are links to the pages:
> http://srcf-ps.stanford.edu/maddash-webui/index.cgi

On this one do you have the perl-perfSONAR_PS-Nagios package installed? and 
do the files /opt/perfsonar_ps/nagios/bin/check_owdelay.pl and do 
/opt/perfsonar_ps/nagios/bin/check_throughput.pl and are they executable? 
What happens when you run "/opt/perfsonar_ps/nagios/bin/check_throughput.pl 
-u http://float10g.stanford.edu:8085/perfSONAR_PS/services/pSB-w 0 -c 0 -r 
900 -l -p" by hand?


> http://rcf-ps.stanford.edu/maddash-webui/index.cgi?dashboard=Stanford%20PS%20Sites

The commands are running fine its just not getting data except for from 
psych-ps.stanford.edu. The problems are summarized by host below:

psych-ps.stanford.edu: MA is running and has throughput data for 3 of the 4 
hosts. The one it doesn't must either not be configured or bwctl is failing 
to complete for some reason. The OWAMP tests are orange because OWAMP regular 
testing is not running. This is most often caused by no OWAMP tests being 
configured.

netspeed1.stanford.edu: MA is running but you are getting no results. This 
host is running v3.2.0 of the toolkit which is quite old. As of March 28th of 
this year we no longer support pre-3.3 versions of the toolkit so i would 
recommend upgrading. Instructions here: 
http://code.google.com/p/perfsonar-ps/wiki/pSPerformanceToolkit33UpgradeGuide

float10g.stanford.edu, rcf-ps.stanford.edu, and srcf-ps.stanford.edu: The MA 
Is not running. Might want to look in 
/var/log/perfsonar_ps/perfsonarbuoy_ma.log. Might be worth just trying a 
reboot too. 

Also, if you want the "Stanford PS Sites Dashboard" to appear by default when 
you visit http://rcf-ps.stanford.edu/maddash-webui/ you can set the 
defaultDashboard to "Stanford PS Sites" as detailed here:
http://code.google.com/p/perfsonar-ps/wiki/MaDDashInstall#Customization

> http://float10g.stanford.edu/maddash-webui/index.cgi

Similar to #1 though the error message is slightly different. It might be 
worth doing a "yum reinstall perl-perfSONAR_PS-Nagios". Something seems to be 
wrong with the commands under /opt/perfsonar_ps/nagios/bin.

> 
> I'm not sure why that one host, psych-ps, does show values on a single host.
> 
> Further thoughts or suggestions?

As a general comment it looks like you have the same dashboards running on 
three different hosts. You might save yourself some headaches if you just 
consolidated to one dashboard to worry about managing. Maybe you have other 
plans for these hosts, but right now they are all attempting to do the exact 
same thing.  I'd recommend consolidating on rcf-ps.stanford.edu since it is 
the closest to working and the issues its reporting appear to be with the 
testers, not the dashboard configuration. 


Hope that helps,
Andy




> 




> Phil
> 
> On 4/23/14, 6:47 AM, Andrew Lake wrote:
>> Hi,
>> 
>> The FAQ here should have the steps you need: 
>> http://psps.perfsonar.net/toolkit/FAQs.html#Q71
>> 
>> Thanks,
>> Andy
>> 
>> 
>> On Apr 22, 2014, at 2:34 PM, Phil Reese 
>> <>
>>  wrote:
>> 
>>> Hi,
>>> 
>>> I'd been using the Dashboard for quite awhile then a disk crash had me 
>>> starting from scratch again.
>>> 
>>> Just got back to it and decided I'd just install it on a different, newer 
>>> system.
>>> 
>>> I did a yum update for all the perfsonar stuff and then did the yum 
>>> install maddash
>>> 
>>> I edited the maddash-server.yaml file as suggested and as I had done in 
>>> the past (didn't have a copy of the old file, of course).
>>> 
>>> Started up the process and the Dashboard shows up just fine but all the 
>>> boxes are orange and don't seem to change over a couple of hours of time.
>>> 
>>> Thought I must have fat fingered something so went back to the original 
>>> system and went through the same steps.  This time the banner appears but 
>>> only a spinning wheel showed, for a long time. Looked in the logs and it 
>>> said:
>>> level=INFO ts=2014-02-05T03:56:25.844168Z event=maddash.init.start 
>>> guid=775ea81a-7665-4590-b159-00612b5a5b6e
>>> level=ERROR ts=2014-02-05T03:56:54.189207Z event=maddash.init.end 
>>> guid=775ea81a-7665-4590-b159-00612b5a5b6e status=-1 msg="Error loading 
>>> database: Column 'TEMPLATENAME' is either not in any table in the FROM 
>>> list or appears within a join specification and is outside the scope of 
>>> the join specification or appears in a HAVING clause and is not in the 
>>> GROUP BY list. If this is a CREATE or ALTER TABLE  statement then 
>>> 'TEMPLATENAME' is not a column in the target table."
>>> 
>>> Is there a way to clean out things and start over?
>>> 
>>> Stilling thinking fat finger issue, I went to my last production PS box 
>>> and did the same steps as the first system and got the same results.  The 
>>> grids show nicely but all have orange boxes that don't change over time.
>>> 
>>> Suggestions?
>>> 
>>> Phil
>>> 
>