Rolf, Jim, Dave

Short story, we installed RCG2.8, we found problems which could not be fixed quickly, we backed out to 2.7.2

The following systems were not changed during the 2.8 testing: h1psl front end, the DAQ code, mx_stream, mbuf.

All front end models were recompiled against 2.8. Local changes were found in h1hpibs, h1hpiitmx, h1isiham6, h1susitmy (the HEPI models showed major changes compared to SVN).

All front end models except the PSL were restarted with the new 2.8 versions. The DAQ was reconfigured to remove all the EDCU INI files relating to front end channels (only non-FE EDCU channels were being acquired, e.g. Vacuum and FMCS).

We found that running the h1iopsush2a model caused corruption of the DAQ data stream from the front ends to the DAQ Data Concentrator. Rolf tried various tests but no quick solution was found. With time running out we decided to back out the 2.8 models and reinstall the tag2.7.2 models.

I ran an "install world" make in the rtbuild-2.7.2 area, so models were not recompiled against 2.7.2, existing binary code was re-installed in the target area.

We ran into safe.snap restoration for the following models: h1hpietmy, h1isibs, h1isietmx, h1isietmy, h1isiitmy, h1pemey, h1peml0. In these cases they were manually restored to this morning's settings before the models were stopped.

To test RCG2.8 we recompiled dataviewer, the new version is still installed as it is backwardly compatible.

EDCU for Beckhoff C1PLC2 was edited to remove two channels (H1:LSC-REFL_SUM_A,B_OUTSW) which appear to have disappeared.