Zabbix alerting going nuts after slave sends data

[qix]

Hello all,

I've setup Zabbix 1.4.4 on two hosts and i wanted to setup distributed monitoring. (node 1 and node 2)
I was smart enough to make a backup of both the databases before setting it up.

The moment I added the master node on the slave and the remote on the master Zabbix started to send alert mails.

These mails seem to originate from a rule I setup on the master node to mail a user when any trigger becomes 'false'. They are also send from the master node.

It looks like this:

On 2008.01.15 14:26:25 the node *UNKNOWN* was cleared of the following problem:

Name: 
Classification: Not classified

If there is more information available about the specified trigger, this will be shown below.
---
(null)

The original message format was:

On {DATE} {TIME} the node {HOSTNAME} was cleared of the following problem:

Name: {TRIGGER.NAME}
Classification: {TRIGGER.SEVERITY}

If there is more information available about the specified trigger, this will be shown below.
---
{TRIGGER.COMMENT}

I get about 55 mails per minute (!)
I already have 5389 in my mailbox

I disabled the rule that says to email this user: no effect
I removed the rule that says to email this user: no effect
I disabled the mail addresa for this user: no effect
I removed the mail address for this user: no effect
I stopped the Zabbix server on the slave: no effect
I stopped the Zabbix server on the master: no more mail

Ok, now what??

I restored the database on the master node (slave node was running again) first and restarted the Zabbix server, very soon the problem popped up again.

Then I restored the database on the slave node (master node was still running) also without effect.

I stopped the slave and master node again and re-restored the databases.
Now without having one the nodes running
I started Zabbix on both the nodes when both the databases were restored.

It seems to run ok now, no more Zabbix spam for now.

This puzzles me the most: Why is zabbix still acting beserk when the database has been restored??
Did the master node get some data from the slave node that made it go beserk? Could it be possible to jkust send some data to Zabbix server and it starts acting all weird like this?

Alexei, I think this is something somebody really needs to look at.
I made a copy of the databases before restoring, so if that is needed, you can get that from me.

I would happily answer any other questions about this problem aswell.