J. Kissel, J. Batch,

Jim informed me this morning that other front-end models, seemingly unconnected to my h1isietmx model restart, had their framebuilder data corrupted immediately following my restart. Hence all frame data from these computers has been hosed since 2014-07-31 22:45 UTC. See Jim's complementary log here.

The corruption is a result of an mxstream glitch sent out by h1isietmx upon restart, that was sent across the ethernet connection to other computers on the same ethernet network (i.e. h1sush2b, h1oaf, h1pemmy, h1pemmx, h1susey and h1lsc). This is a known problem at least by Jim (who will subsequently write a bugzilla report).

To identify the problem:
- Open CDS_OVERVIEW screen
- Look for models / computers that have *both* the DAQ (Data AQuisition system) and the DDC (Daq Data Concentrator) showing red status in their CDS STATE WORD, (e.g H1:FEC-103_STATE_WORD).

To solve the problem:
- Log into corrupted front-end
- run /etc/start_streamers.sh
This reboots the mxstream process, occasionally glitches the framewriting PSL front end (?!), presumably fixes the problem, and restores the STATE WORD to green.