Problem with distributed monitoring (nodes)

[marakshin]

Good day.
In the past, we have a 2 separate zabbix servers. In april 2014 we want to use distributed monitoring with nodes, but have a strange problems with sync:
1) in child node, node watcher is busy (100%)
2) in log file on the master node, we seen a message:

NODE 1: cannot receive answer from Node [2]: ZBX_TCP_READ() failed: [4] Interrupted system call

3) in log file on the child node, we seen a message:

NODE 2: sending alerts of node 2 to node 1 datalen 9282045
NOT OK

All servers is Zabbix 2.2.3 version, Mysql server queue (show processlist is not too high...
Anybody have a same problems? What I can debug?

Thanks!

[esawyja]

Hi,
We also use distributed monitoring, but the support for it seems to be limited. I have recently switched over to this method, that seems to work for me, see below

http://kingratlinux.blogspot.com/2014/07/solution-to-bypass-distributed.html