Date: Sat, 11 Sep 2021 11:17:29 -0600
Subject: Re: [sloan-25mlog 5647] 2.5m obslog 20210910 (SJD 59468)
From: "Daniel Oravetz"
> * We had another instance where the MCP became unavailable, but we
> troubleshooted
>  the problem and now have a solution on how to safely fix this
> communication issue between systems.  See log for details.
For clarification, this doesn't look like the same kind of "MCP
unavailable" problem that happened on August 25th - there aren't any
dropouts in the MCP logs, for example. This seems to have been a
communications jam-up after the TPM restart.
The GUI was displaying some info correctly (e.g., MIG values that come
from a serial port), but it wasn't displaying the MCP data that comes from
the TCP/IP MCP data packet even though the MCP was working fine in the
lead-up to the TPM restart. When the GUI isn't displaying the MCP info it
shows "MCP unavailable", but that doesn't necessarily mean the MCP isn't
supplying the data on it's side, it's just not getting through.
As an analogy, if one opens an MCP/TPM GUI, but their local network isn't
allowing the TPM multicast datagrams through, then in the upper-right
corner there will be a "TPM unavailable" message. That doesn't mean the
TPM itself is offline or having problems, it means that instance of the
GUI isn't getting the datagram.
Anyhow, seemed worth clarifying that this doesn't look like a repeat of
whatever happened on August 25th - "unavailable" messages can have
different causes.
> Observers: next time you see the weather status page freeze for the 2.5m
> weather telemetry, please just restart the *optidew* (there is a wiki page
> with the instructions).
One more useful thing to try is restarting the 'humid' task. The 'humid'
task connects to the humidity sensor at the terminal server. It actually
reads the humidity-sensor data and populates the datagram. So if
restarting the 'optidew' doesn't result in the weather data refreshing on
the Weather Status page, try restarting the 'humid' task as well. You do
it the same way as you would for 'optidew' but with 'humid' instead:
./proc -stop humid and then ./proc -start humid.