zabbix 1.6.4 DM Node missing in frontend

[ataylo13]

I am testing the latest stable 1.6.4 frontend. I noticed that one of my nodes is missing from the interface and drop downs. I have 2 Child Nodes and 1 proxy.

NodeId 2 was converted to proxy
NodeId 3 is coming in fine
NodeId 4 is not in drop down


From the 1.6.2 frontend
All nodes are there, data is being collected, etc.

What other information do you need to help troubleshoot this one?

[webdisk]

nobody knows

I have similar problem.
the nodes existing and working, but I cannot see them in master node, even the master sending out ntification on behalf of child node.

it seem 1.6.4 bug?

[webdisk]

zabbix 1.6.4 DM Node missing in frontend, while 1.6.3 DM ok

Description of the node missing issue:

I've upgrade master (ID8)and 2 child node (ID10 & ID20) from 1.6.3 to 1.6.4, in webpage "administration/node":

1. master node display ID8 & ID10, missing ID20;
2. child nod ID 10, 20 display self only, missing ID8;
3. @child node, copy back 1.6.3 frontend, master Node displayed;
4. @master node, copy back 1.6.3 frontend, all Nodes displayed correctly;

It's need to be fixed asap as now I'm running 1.6.4 binary with 1.6.3 php.

Thanks in advance to development team.

[Aly]

Could you describe your node configuration more in structured way please. So I could build same setup to test it, due to in my current setup I don't have any problems with displaying nodes.

[webdisk]

description for the missing node scenario

thanks a lot for prompt reply.

1. zabbix runnung over centos5.2, MYSQL5.0.25.
2. all nodes running over vmware server; node 8 has 2 children: ID10,20. all nodes listen tcp:10051.

3. master node ID8: zabbix 1.6.2 upgraded to 1.6.3, tnen, to 1.6.4.

4. child node ID10: same as master node

5. child node ID20: FAN CD1.1, nagios/centreon, centos 5.2, then I installed zabbix over it. so hr mysql has multiple database.

6. I have tried mysqldump to backup configuration on master node and node 20.

Gain confident in zabbix and ask our colleague for comercial supprt from zabbix.

hope I state it clear enough.

[wax66]

Node not appearing in dropdown list.

Same issue here, where my #3 node does not show in dropdown on the web interface, similar setup as above.

CentOS 5.2, mysql 5.0.45, went from 1.6.2 to 1.6.3 to 1.6.4, all source compiles and installs, x86_64, php frontend. All 3 of my nodes are pretty much identical. 1 master, 2 children, no proxies.

According to the logs it does appear to still be talking, just not showing in the UI.

6623:20090409:083041 NODE 3: Received data from master node 1 for node 3 datalen 149
6623:20090409:083042 NODE 3: Sending history_sync of node 3 to node 1 datalen 5417
6623:20090409:083042 NODE 3: Sending history_uint_sync of node 3 to node 1 datalen 1561

-Ron

[xs-]

probably related to

http://www.zabbix.com/forum/showthread.php?t=12162

[wax66]

Possibly, but I am logged in as a Super-Admin, I only have 1 master node, the one I'm logging into, and only one child node out of the 2 is showing.
-Ron

[Aly]

Honestly, I can not find any reason why some nodes aren't in the dropdown, repeat the problem I can't either. There are some issues fixed related to nodes, may be that one is fixed within.

Changes in rev. 7173. You may try downloading the nightly build from http://www.zabbix.com/developers.php when it will be available.

[wax66]

I'm kinda stumped too. Even worse, there are other UI issues I've come across like some templates not showing in certain areas and such, but I thought I'd take a screenshot of another UI issue I'm having which is more annoying, and probably related to my node dropdown issue.

The screenshot is of the "Group" dropdown in Configuration->Hosts. As you can see, they are all 'monitor-dc1', whereas the data in the database appears fine:


mysql> select * from groups;
+-----------------+------------------------+
| groupid | name |
+-----------------+------------------------+
| 300100000000002 | BigIP Servers |
| 100100000000005 | Infrastructure Servers |
| 200100000000001 | Infrastructure Servers |
| 300100000000001 | Infrastructure Servers |
| 100100000000002 | Linux servers |
| 200200000000002 | Linux servers |
| 300300000000002 | Linux servers |
| 100100000000006 | Network Devices |
| 300100000000003 | Network Devices |
| 100100000000001 | Templates |
| 200200000000001 | Templates |
| 300300000000001 | Templates |
| 100100000000003 | Windows servers |
| 200200000000003 | Windows servers |
| 300300000000003 | Windows servers |
| 100100000000004 | ZABBIX Servers |
| 200200000000004 | ZABBIX Servers |
| 300300000000004 | ZABBIX Servers |
+-----------------+------------------------+
18 rows in set (0.00 sec)

[Aly]

I'm kinda stumped too. Even worse, there are other UI issues I've come across like some templates not showing in certain areas and such, but I thought I'd take a screenshot of another UI issue I'm having which is more annoying, and probably related to my node dropdown issue.

The screenshot is of the "Group" dropdown in Configuration->Hosts. As you can see, they are all 'monitor-dc1', whereas the data in the database appears fine:


mysql> select * from groups;
+-----------------+------------------------+
| groupid | name |
+-----------------+------------------------+
| 300100000000002 | BigIP Servers |
| 100100000000005 | Infrastructure Servers |
| 200100000000001 | Infrastructure Servers |
| 300100000000001 | Infrastructure Servers |
| 100100000000002 | Linux servers |
| 200200000000002 | Linux servers |
| 300300000000002 | Linux servers |
| 100100000000006 | Network Devices |
| 300100000000003 | Network Devices |
| 100100000000001 | Templates |
| 200200000000001 | Templates |
| 300300000000001 | Templates |
| 100100000000003 | Windows servers |
| 200200000000003 | Windows servers |
| 300300000000003 | Windows servers |
| 100100000000004 | ZABBIX Servers |
| 200200000000004 | ZABBIX Servers |
| 300300000000004 | ZABBIX Servers |
+-----------------+------------------------+
18 rows in set (0.00 sec)

Known problem and there is a patch for it:

http://www.zabbix.com/forum/showthread.php?t=12244

[ataylo13]

I just downloaded the latest developer release (7173) and I am now missing one of my nodes, Node 4.

Looking at the logs I am still receiving the data from Node 4 and when I connect directly too it, the server seems to be running fine. When I go to the Administration->Nodes tab, Node 4 is not there on the Master.

I still have the 1.6.2 frontend in production, and it is showing all the nodes with their correct alerts, etc.

[wax66]

Further info

That patch worked great for correcting the group display, thanks!

Additional info probably related to this issue:

The missing node (node 3) doesn't show in any of the following pieces of the UI on node 1 (the master node):

Administration -> Nodes
Monitoring -> Dashboard
Monitoring -> Latest Data/Triggers/Events -> Group pulldown

No alerts that show up on node 3 are shown in node 1.

The Administration -> Nodes on node 3 only shows itself (as if I had never added node 1).

But mysql knows the real story:
mysql> select * from nodes;
+--------+-------------+----------+--------------+-------+---------------+--------------+----------+----------+
| nodeid | name | timezone | ip | port | slave_history | slave_trends | nodetype | masterid |
+--------+-------------+----------+--------------+-------+---------------+--------------+----------+----------+
| 1 | monitor-dc1 | -5 | 172.22.32.99 | 10051 | 90 | 365 | 0 | 0 |
| 3 | monitor-dc3 | -5 | 127.0.0.1 | 10051 | 30 | 365 | 1 | 1 |
+--------+-------------+----------+--------------+-------+---------------+--------------+----------+----------+

As does mysql on node 1:
mysql> select * from nodes;
+--------+-------------+----------+---------------+-------+---------------+--------------+----------+----------+
| nodeid | name | timezone | ip | port | slave_history | slave_trends | nodetype | masterid |
+--------+-------------+----------+---------------+-------+---------------+--------------+----------+----------+
| 1 | monitor-dc1 | -5 | 127.0.0.1 | 10051 | 30 | 365 | 1 | 0 |
| 2 | monitor-dc2 | -5 | 172.20.32.103 | 10051 | 90 | 365 | 0 | 1 |
| 3 | monitor-dc3 | -5 | 172.20.32.66 | 10051 | 90 | 365 | 0 | 1 |
+--------+-------------+----------+---------------+-------+---------------+--------------+----------+----------+

It's almost as if node 3 is just completely orphaned. But in the logs I still see the sending of history syncs to node 1.

Any ideas of troubleshooting I can do? I can't do much to node 1 and node 2, those are pretty high importance, but node 3 is in a lab where I can experiment a bit.
-Ron

[Aly]

Try to comment in file include/config.inc.php line ~263:

$ZBX_NODES = get_tree_by_parentid($ZBX_LOCALNODEID,$ZBX_NODES,'masterid'); 

to:

//$ZBX_NODES = get_tree_by_parentid($ZBX_LOCALNODEID,$ZBX_NODES,'masterid'); 

If that doesn't helps, add before that line this piece of code:

 SDI($ZBX_NODES); 

You will see the debug info on any page possibly with some errors.. Copy that debug info here please, it will help me to find the reason.

[wax66]

Et voila!

Yup, commenting out that line fixed it. Only changed it on the master server and just to be clean I restarted httpd, and sure enough now node 3 is shown everywhere!

Thank you very much!

Should I worry about that affecting anything negatively?

And I'm assuming I should go ahead and do that on the 2 child servers (not that I use the UI there much, if it all).
-Ron