I hadn't realized that Dave and Jim B were already on the case for getting the CW injections to work again.  Dave emailed today: "Jim and myself have been installing Keith’s LLO modifications to get monit to autostart the psinject process. We are not quite there for the autostart, but we were able to run psinject manually as user hinj. I’ll take a look at why your manual start yesterday did not work.  Once we have psinject running under monit, we’ll do the same for tinj."

I added some debugging lines to the x_exec_psinject_from_script script to hopefully help track this down.  It's clear that there is a difference in environment variables when trying to start psinject from the start_psinject script versus just executing psinject in an interactive session.  You can compare, for instance, the file Details/log/x_exec_psinject_from_script.log against Details/log/interactive.log, and there are substantial differences, such as:

[hinj@h1hwinj1 Details]$ diff log/interactive.log log/x_exec_psinject_from_script.log
...
7c9,10
< APPSROOT=/ligo/apps/sl6
---
> APPSROOT=/ligo/apps/linux-x86_64
> ASC=H1:ASC
11a15
> CDSDIR=/ligo/cds/lho/h1
...
27c33,34
< EPICS_BASE=/ligo/apps/sl6/epics-3.14.12.2_long/base
---
> EPICS_BASE=/ligo/apps/linux-x86_64/epics/base
> EPICSBIN=/ligo/apps/linux-x86_64/epics/base/bin/linux-x86_64:/ligo/apps/linux-x86_64/epics/extensions/bin/linux-x86_64
...

Evidently, when executed from within the script, the environment is simply not set up properly to allow psinject to run -- e.g., libCore.so is not being found.  I'll let Dave and Jim take it from here.