Grafana on Cornell apollo sends some data to wrong $SN

[pwittich]

(As discussed with @dgastler )

I have a blade where the Grafana data is being split between blade 00 and blade 01 (where 01 is the right value). The firefly data is the only one that’s wrong, the other all go to the right place.

This shows the uptime for what is logged as Apollo00 and Apollo01 and one Firefly channel. you can see by the changes of colors two transitions.

• when the MCU uptime goes from one color to the other, the grafana lost the $SN and all gets sent to 00.
• when the Mcu uptime goes back to the right place after a cold reset, note that the FF does not change color back from green to yellow, until later. there the MCU uptime does not change color again.

(I scaled the MCU uptime to fit it on the plot by 40/1000)

The last transition (where the FF data goes to the right place) is associated with a restart of the graphite_monitor service on the blade.

[pwittich]

Seems to happen when the IPMC reboots the blade