Originally posted by vasiliy
-
The release has been carefully tested, especially everything related to distributed monitoring. I do not believe the configuration exchange does not work... -
Alexej, what am I missing?Originally posted by Alexei
You still want us to do the following:
These are the steps I followed:
On CentralNode.
1.) drop the database, recreate a new db, and reimport your initial database files (is this sufficient for reinstall? should i do anything else?)
2.) vi /etc/zabbix/zabbix_server.conf... The node_id is already set to 1.
3.) ./zabbix_server -n 1 -c /etc/zabbix/zabbix_server.conf
this script reported that tables have been converted.
4.) navigate to http://zabbix1/zabbix/, login as admin, now i see a nodes tab.
5.) click on the current node that i see, change its name to CentralNode, and change its IP from 127.0.0.1 to 10.0.0.0.4
6.) click add new node, remote, call it Child.1, put in the ip address... done.
What I see at the end:
1 /CentralNode Local GMT-04:00 10.0.0.4:10051
2 /CentralNode/Child.1 Remote GMT-04:00 10.0.0.5:10051
7.) havent started the service yet.
On Child.1.
1.) drop the db, recreate, yada yada
2.) make sure /etc/zabbix/zabbix_server.conf's Node_id is =2
3.) ./zabbix_server -n 1 -c /etc/zabbix/zabbix_server.conf
4.) navigate to http://zabbix2/zabbix/, login as admin, click Nodes.
5.) Here I change the name of the current node I see from localnode to Child.1, change ip from 127.0.0.1 to 10.0.0.5.... Add a master node....
After doing this.... I see this as 1 line...
2 /CentralNode/Child.1 Local GMT-04:00 10.0.0.5:10051
should I see anything else?
now, i fire up the services on both machines....
When I make configuration changes, or add hosts vi the web interface on CentralNode, the child does not receive them.
THANKSComment
-
Each node must contain information about itself and all child and master node. You should have at least two nodes defined there.Originally posted by vasiliyComment
-
I did that....Originally posted by Alexei
Its certainly not letting me create master...
I will restart for the 100th timeComment
-
Alexej, I was correct...
Here are my screen shots:
On central Node:
On child node:
Comment
-
As you can see, after adding a master node to the child... it only displays node 2(itself) underneath the master.
Is this correct? Am I following things properly?Comment
-
Forgive me if I am wrong, because I have not set up DM myself.
I am in the process of setting it up now that 1.4.2 is officially released. However, I noticed a difference between your procedure and the manual.
On your Central Node you start up withOn the Child Node you start up using the same node number as the Central Node.Code:3.) ./zabbix_server -n 1 -c /etc/zabbix/zabbix_server.conf
The manual says it should be started withCode:3.) ./zabbix_server -n 1 -c /etc/zabbix/zabbix_server.conf
Is that just a typo on the forum, or is the manual incorrect?Code:./zabbix_server -n 2 -c /etc/zabbix/zabbix_server.conf
What confuses me is that in section 12.3.2 of the manual the child should be started with a -n <node_id>, but in section 12.5 it says a child can be started with a simple ./zabbix_server with no node id specified.
Alexei can you tell us if the -n <node_id> is required when starting a child node? or is it only required the first time a child node is started but can be left off at any other time?
Thanx.Comment
-
The -n option is not for server startup! It has to be executed only once (!) after the database is created and default data (data.sql and images) is loaded.Comment
-
that is correct, it was a typo on the forum.Originally posted by Alexei
so whats the suggestion? am I missing something?Comment
-
Yes, the master node details must present in configuration of the child node. I am not quite sure why you cannot add it...Originally posted by vasiliyComment
-
Alexej, i have added it. When I try adding it again, it says that master node exists.Originally posted by Alexei
please help!
I follow your directions to the T.
when I add a new master node to the child, it replaces the local node layout... like so:
LocalNode gets replaced with Masternode/Localnode
what am i missing!Comment
-
Hi,
I just upgraded to 1.4.2 and DM is still not working.
Even when i do a tcpdump on my slave, i don't see any connections coming from the master.
Is there a howto online for DM? I got all my information from this forum, but maybe i am making a mistake?Comment
-
Wido, I am having the same issue...
I really do not know what we are doing wrong!Comment
-
In your case, the child node must have exaclty the same configuration as the master node. Two nodes were successfully created for Node 1, so you may repeat exactly the same steps to create the two nodes for Node 2.Originally posted by vasiliyAlexej, I was correct...
Here are my screen shots:
On central Node:
On child node:
I have a test system which is configured in exactly the same way and its nodes operate absolutely fine.Comment
-
I did a mistake by importing data first into the DB (zabbix_server was not running) and then configuring the node id.Originally posted by Alexei
The zabbix_server tries to converts the data, but it doesn't do it
completely.
So the statement of Alexei quoted above is completely valid.Code:41653:20070823:163058 No function for functionid:[1000000000011999] 41653:20070823:163058 Expression {500000000011999}>0&{1000000000011999}=0] cannot be evaluated [Unable to get value for functionid [1000000000011999]]
Everything else is "walking on thin ice".
It might work, e.g., depending on your triggers defined.
I leave the decision to Alexei whether this need to be fixed or not.
I did learn from this experience and it was easy to fix, anyway.
Details:
Export was done on 1.3.8, import on 1.4.2
I did change the nodeid on the 1.3.8 server from 0 to 1 - and because
of a bug in the Postgresql schema - back to 0.
So this problem might be fixed in the current release or even introduced
by the 1.3.8 beta. I don't know.
Regards,
Norbert.Comment
Comment