Hi Ton, thanks for the reply.

Here are the results of that query:

mysql> select count(*) from nagios_conninfo where instance_id != 2;
+----------+
| count(*) |
+----------+
|        0 | 
+----------+
1 row in set (1.15 sec)

I am running the databases on a different server than the Opsview master.
So it¹s possible that the NDO query for the default instance id fails
occasionally due to something in the network.

I did go through and make all the MySQL tuning adjustments last week.
However I haven¹t restarted MySQL yet so they haven¹t taken affect.

Mike


Mike Dorman
Lead Systems Engineer

Latisys-Denver, LLC
(303) 268-1504 ­ direct
[email protected]
www.latisys.com



From: Ton Voon <[email protected]>
Reply-To: Opsview Users <[email protected]>
Date: Fri, 23 Oct 2009 09:38:23 +0100
To: Opsview Users <[email protected]>
Subject: Re: [opsview-users] Opsview 3.3.2 upgrade warning


On 20 Oct 2009, at 18:22, Mike Dorman wrote:

>  I just did the check for this bug on our system and the minimum instance ID
> is 2, so looks like we are affected.  In looking at the FAQ, it says that
> there should only be one row in the nagios_instances table.  However, we have
> 24 (see below.)

It is very strange that the id numbers increment by 2 each time.

> I am not very familiar with NDO, but it seems like this might be indicative of
> a larger problem.  The FAQ says that ³you have more than just Opsview pointing
> to this runtime database², but I know that is not true.

The problem is in NDO code. It queries the nagios_instances table for the
"default" instance to get the instance id. For some reason which we haven't
determined yet, it sometimes doesn't get the existing row. Thus the next
step it takes is to insert a new row and then use that new id number.

What we did was to set a constraint on the table so you can never have two
instance ids with the same name. Then we changed the NDO code to select
again after attempting the insert. This should stop what you are seeing
below in new systems.


> Do I need to worry about this?  Or just wait for v3.3.3 and everything will be
> fine?  Right now we are still on 2.14.1 (I know, I know) and I am getting
> everything in place to go to the latest v3.x.

The only symptoms we are seeing are duplicate numbers of services in the HH
view. Can you try this statement to see how many times you can getting
non-instance id = 2 connecting:

select count(*) from nagios_conninfo where instance_id != 2

Also, we seen that tuning mysql will help the database:
http://docs.opsview.org/doku.php?id=opsview3.3:mysql

Ton



_______________________________________________
Opsview-users mailing list
[email protected]
http://lists.opsview.org/lists/listinfo/opsview-users

<<image.png>>

_______________________________________________
Opsview-users mailing list
[email protected]
http://lists.opsview.org/lists/listinfo/opsview-users

Reply via email to