Vivek Khera wrote:
> On Aug 21, 2007, at 7:56 PM, Michael Best wrote:
>
>   
>> The plugin as it works now, only appears to report if the slony  
>> messages
>> are getting through, not so much that the data is in sync, I could be
>> wrong on that, but it reports OK for both of my hosts even when one of
>> them is quite out of sync.
>>     
>
> Once your data is big, and your cluster gets bigger, you will get a  
> trigger nearly every time you run the script, then.  It is rare for  
> all of our nodes to be totally in sync at any given time.  usually  
> one is at least 1 event behind.
>
> With your modification, we'd have alarms going off all the time.
>
> You want a separate test to determine if your replication delay is  
> tolerable and trigger when it is out of tolerance.  This is  
> asynchronous replication, so it is not an error to be out of sync.

Okay, thanks for the clarification, I was questioning the premise of the 
test, good to hear that the normal nagios test works at some level.

I was having problems with slony where my nodes weren't getting data at 
all due to an error, and I was questioning the usefulness of the nagios 
plugin when it wasn't even alerting me to the fact.

My modification will likely be useful for this particular setup.

-Mike
_______________________________________________
Slony1-general mailing list
[email protected]
http://gborg.postgresql.org/mailman/listinfo/slony1-general

Reply via email to