Thread not up alerts

[Alfred Cox]

We are on CL 6.1.2 and trying to set up alerts for the threads on a process so that if any of them go to a status other than up, the alert will bounce the thread. I can do this for any 1 thread, what I am looking for is a way to do this for all the threads on a process where it only bounces the thread that isn’t up but i don’t have to list a new alert for every individual thread. Is that possible?

Topic

[Peter Heggie]

I think you need an alert that runs a tcl which will detect this scenario and then fire a tcl action to recycle the identified threads.

Peter Heggie
PeterHeggie@crouse.org

[Michael Hertel]

Wouldn’t you configure the “protocol status”, select “source” of the threads of interest, then select a “source count” of “any”?

[Alfred Cox]

The trigger I have set up just like that. the problem is finding the variable in the action so that it only restarts the thread that isn’t up instead of all of them.

hcicmd -p epic_adt_test -c ” prestart 1″

[Michael Hertel]

We pass %A to a tcl script.

In a portion of the script we do the following:

#######################

# Protocol Status

#######################

       if {[string equal [string range $alert 0 21] “Thread protocol status”]} {

               set threadname [lindex [split $alert ” “] [lsearch -regexp [split $alert ” “] _]]

               msiGetStatSample $threadname inf

               set qd1 UNKNOWN

               keylget inf PSTATUS qd1

               set alert “Protocol Status: $qd1”

               set sub “$threadname Protocol Status: $qd1”

               if {[string equal $qd1 initializing]} {return}

       }

Perhaps this can lead you in the right direction.

[Author]

Replies