INI sometimes hangs

[jaycedowell]

I've ran into this a couple of times at LWA1 now where I send an INI and then it get stuck somewhere along the process. Today it was after the "FPGA configuration validation" step, or at least that was the last INFO update.

[jaycedowell]

Possibly related: Yesterday I was looking at the LWA1 OpScreen page and I noticed that the last health check was ~25 hours ago. Could the stuck INI be related to another lockup with the ZCUs? We did go through this before (but I can't find where I documented that problem).

[jaycedowell]

Happening right now. I don't see anything holding a ZCU102 lock.

Update: But I do see something odd in the MSC logs. <class 'OSError'>: [Errno 24] Too many open files: '/tmp/260428_000400_snapspecs.dat'| in response to a RPT HEALTH_CHECK.

[jaycedowell]

After some digging it looks like the problem is related to 77a6693 and how we throw away the self._zcu reference to trigger a new connection.

[jaycedowell]

Rolled out for testing this morning.

[jaycedowell]

This really needs some testing. Maybe after we get things settled with #44.