CentOS 6.x, galera cluster, mysql5.6

This does not look like a bug report and it might be better not to try to solve your problem here. Rather, please consider https://www.zabbix.org/wiki/Getting_help for ways to get support from the community.

These triggers may go into an unknown state for many valid reasons. For instance, you have restarted your server recently and nodata() cannot be evaluated properly (see ZBX-7628). Alternatively, these items may not have data in them to evaluate str() function.

The only thing I am suspicious about is that all your triggers seem to reference logrt[] items, but the error message complains about log[] item. Could you please show the full trigger expression that fails to evaluate with this error message?

PS: Note that you might or might not have a typo in "robot but is in pending REQUETS".

I had changed the the log back to logrt, my exact point, the expression is not picking up the changes being made in the trigger. Instead of updating it goes into an unknown status. I tried to edit previous triggers that had gone into an unknown status by making expression simpler, ie. removing nodata check, and it stayed at unknown and did not evaluate new expression. If I clone a trigger that has gone unknown, and change the name of the Trigger, then it repopulates correctly. I also tried to add new triggers with and without nodata values, and they populated correctly. The issue is that if you try to fix an unknown trigger and change the item, the unknown triggers are not being re-evaluated.

Example 1 - trigger with unknown status and evaluation error
Netbackup Syslog Solaris Tape in pending requests message Sev 3

=1 &

=0

Example 2 - trigger which was cloned and renamed and is now in an enabled status
Syslog Solaris Netbackup Tape in pending requests message Sev 3

=1 &

=0

The problem does not reproduce currently with Zabbix 2.2.0.

I create a character item that has no data. Then I create a trigger with str() and nodata() functions, which correctly becomes unknown, because there is no data for this item and str() fails to evaluate. Then I clone that trigger and the copy becomes unknown after a while, too, for the same reason. Then I send some data in and the triggers turn into a normal state.

Could you please describe how to reproduce the problem reliably? In "Monitoring" -> "Latest data", do you see any values for your item?

Issue being seen with nodata will be resolved in 2.2.3

given that we are not aware of specific changes to fix this, reopening to set to "cannot reproduce"