That is quite easy to do.

Here is the lua code for that bit of the script…

      controller = r_value.addr.."/controller"

            print ("  Restarting robot: " .. r_value.name)

            out,rc = nimbus.request(controller,"_restart")

            if rc == 0 then

               print("  Robot:" .. r_value.name .. " : restart executed \n")

            else

               print("  ******* Robot:" .. r_value.name .. " : restart failed \n")

            end

you just need to setup the r_value.addr which is the fully qualified robot address (/domain/hub/robot)

cheers

Ok I'm feeling kind and will give you the script

Try this script in your AO profile…

You can test it by putting in a nimid of an existing alarm in the alarm get (see commented line below)

--

-- robot_restart.lua

--

print('Robot Restart script')

print('====================')

print(' ')

--al=alarm.get("RC43670325-12873")

al=alarm.get()

domain=al.domain

hub=al.hub

robot=al.robot

robotaddr = "/"..domain.."/"..hub.."/"..robot

print ("  Restarting robot: " .. robotaddr)

         controller = robotaddr.."/controller"

            out,rc = nimbus.request(controller,"_stop")

            if rc == 0 then

               print("  Robot:" .. robotaddr .. " : restart executed \n")

            else

               print("  ******* Robot:" .. robotaddr .. " : restart failed \n")

           end

You need to put in your actual current alarm when you run in the editor.

Replace "RC43670325-12873” with one of your nimid's and uncomment the line.

Comment the "al=alarm.get()” out while you are testing;  this line gets the nimid from the autooperator when you run it properly.

When you run without the nimid, the script is being triggered by the auto operator which is using the alarm to trigger it, which is where it gets the nimid from.

If your criteria is wide then the script will be triggered for multiple alarms and therefore restart multiple robots

You can run the script on a schedule and code a loop so that it restarts all robots attached to a particular hub.

There are quite a few different options for you.

Regards

There are a few problems here.

You are mixing arrays and pds's so your loop is not going to work (#al=0)

But my first question is why alarm.list ?

It won't work with the AO profile as only alarm.get() works that way so "al” will be empty.

If this is triggered by an alarm then you need to use alarm.get().

I think you are trying to run this script for all your alarms of "Max restarts….” In which case you don't need the loop.

For loops are only used when, for example, you are processing multiple rows from a database call.

Also I don't see the point of the sleep(100).

Here's is what I would code, if I understand your req's correctly…

al = alarm.get ()

if al ~= nil then

         if al.prid == "controller" then

            if string.match(al.message,"Max. restarts reached for probe") then

            print("message match")

            probe = string.gsub(string.match(al.message,re),"'","")

            print(al.message.."! Probe-> "..probe)

            addr = "/"..al.domain.."/"..al.hub.."/"..al.robot.."/".."controller"

            printf("/"..al.domain.."/"..al.hub.."/"..al.robot.."/".."controller".."<->Probe="..al.prid)

            local args = pds.create()

            pds.putString(args,"name",probe)

            nimbus.request(addr,"_stop",args)

            pds.delete(args)

--sleep (100)

end

end

end

also you proabably should put some error handling after the nimbus.request.

cheers