Log entry time 14:39:39 on April 10, 2011

Entry number 354964

This entry is a followup to: 354951

I came in and noticed we had this situation :

ls -la /adaql2/data*/e07006_3393*
-rw-r--r-- 1 a-onl a-adaq 4145938432 Apr 10 09:04
/adaql2/data2/e07006_3393.dat.0
-rw-r--r-- 1 a-onl a-adaq  170065920 Apr 10 09:07
/adaql2/data3/e07006_3393.dat.1
-rw-r--r-- 1 a-onl a-adaq  971571200 Apr 10 09:29
/adaql2/data4/e07006_3393.dat.0

Ok, this is the disease of CODA 2.6.1 that it sometimes writes the
same run number.  The file with 971571200 bytes was run 3394.  There 
were no other files with run 3394, and the next was 3395.  So I moved
e07006_3393.dat.0 to e07006_3394.dat.0 on /adaql2/data4.

I also noticed that the correct e07006_3393.dat.0 was written to
MSS tapes, and the 2nd "fake" one was not.  This is as it should be.
The scripts that automatically write to MSS will pick up run 3394
shortly.

When this happens (duplication of data files) the shift should try to
notice it, and restart CODA (startcoda).

And yes I will mention it to the DAQ group.  It's a rare problem
and hence difficult to find and fix.  From what I gather, it is not
happening in Hall C, so it may be how CODA interacts with our script
that picks out the next disk for writing, which is unique to Hall A.