IOC: Failed to start (local)

On my dev machine the system tests failed because SIMPLE was not started. Looking in the log I see:

[Fri Oct 11 16:08:25 2019] [1570806505.75] INFO: The requested data transfer is greater than available memory or EPICS_CA_MAX_ARRAY_BYTES
[Fri Oct 11 16:08:25 2019] [1570806505.75] MAJOR: Could not retrieve IOC list: 'NoneType' object has no attribute 'decode'
[Fri Oct 11 16:08:25 2019] [1570806505.85] INFO: User specified timeout on IO operation expired
[Fri Oct 11 16:08:25 2019] [1570806505.85] MAJOR: Could not (re)start IOC SIMPLE: Could not find IOC (TE:NDW1798:CS:PS:SIMPLE)

It appears that if the status of the
`get_ioc_status` can raise an exception (BlockServer/epics/procserv_utils.py line 68)
`_start_config_iocs` can then not start an IOC (block_server.py line 380)

The IOC list is too long for epics to read:
```
C:\Instrument\Apps\EPICS>caget -w 60 -S -t %MYPVPREFIX%CS:BLOCKSERVER:IOCS|uzhex
CA.Client.Exception...............................................
Warning: "The requested data transfer is greater than available memory or EPICS_CA_MAX_ARRAY_BYTES"
Context: "op=0, channel=TE:NDW1798:CS:BLOCKSERVER:IOCS, type=DBR_TIME_CHAR, count=128000, ctx="unable to fit read notify response into server's buffer""
Source File: ../getCopy.cpp line 92
Current Time: Fri Oct 11 2019 16:59:04.399709900
..................................................................
Read operation timed out: some PV data was not read.

C:\Instrument\Apps\EPICS>set EPICS_CA_MAX_ARRAY_BYTES
EPICS_CA_MAX_ARRAY_BYTES=130000
```

Something is wrong but it is last thing on a Friday and I can not sort out what is going on.