Groups and Dependencies issue

From the sound of what you described it would seem like there shouldn't be any additional weird logic, firewall goes down and is the parent of a group then everything in that group gets marked unreachable.  There's not much else to it.  In general with dependencies there is always a bit of leakage when the child nodes get marked as being down before the poller realized the parent is down, but if you had problems with that aspect of things then I would expect you to have it consistently across all groups, not just the firewall group.  This isn't just a matter of timing right?  The firewall children never get marked as unreachable?  How did you build those group definitions, if it was a script maybe there was something that is wrong with that script?  Does the firewall dependency relationship not have the setting correct for the options on how it should handle the parent down status?  You can set that to where the parent actually never does change the status of the children.

Thanks for the quick response.    Actually I set them up by hand.   So here are the steps on how I did it.  

Just using the previously attached word file as a reference:

1. Created group 020231-020232-MO

In that group I selected the following devices.

020231-020232-Fortigate(that is the firewall)

020231-Mobo

020231sccppc

020232 Mobo

020231 ser

020232 ser

020232 sccppc

2. Created group 020231-020232-FW

Contents

020231-Mobo

020231sccppc

020232 Mobo

020231 ser

020232 ser

020232sccppc

3.  Then I created dependency 020231-020232-Rv50

Parent is 0202031-0202032-RV50

child is 020231-020232-MO  (group)

4.  Then I created dependency 020231-020232-Fort

Parent is 020231-020232-Fort

Child is 020231-020232-FW (group)

Both dependencies are set that when parent is down,m the child will be unreachable and alerts on child will be suppressed.

I set this up on every group and dependency identical as above with the correct groups and parents and children.

100% of the time, when the RV50 goes offline- it reports down, and there are no alerts on anything in group -MO.   When I look, i can see that they are all unreachable.

Likewise, 100% of the time, when the Fortigate or ASA or Juniper, whichever firewall we are using on that particular location goes offline, then every member of the -FW group creates an alert that they are down.   They never become unreachable unless the modem goes down at the same time. 

At first I was wondering if the issue was that the nodes in the MO child group were the exact same as the ones in the FW child group with the exception of the Fortigate is in the modem group but is parent of the FW group.   But someone told me it would make no difference.