Dennis, 

There are still problems with the startup MSEED_Archiver timeseries data.

I am concerned that the continuous MiniSEED archiver 
is too far removed from the digitizer data source
in the chain of data flow.  From what I can see,
the data path for the 2 MiniSEED data streams are:
a.  digitizer -> RockToSLlink
b.  digitizer -> Data_Integrator -> Continuous_Recorder -> MSEED_Archiver

In particular, it appears that the Data_Integrator may 
be responsible for inserting the ~ 0.5160 seconds of gap-filling
flat line data into the time series (see below).
It there any way to suppress this operation?

This test done using:

Using rock 3.7.4i BETA until 2011/01/20,01:40 UTC
Using rock 3.8.1b BETA after 2011/01/20,01:40 UTC

"seed" data -  from rock MiniSEED archiver.
"slink" data - from RockToSLink server.

1.  Before new software installation and reboot:
slink and seed data have same timestamps.
Transition from 100 Hz to 1 Hz is every 30 seconds at ~
	xx:xx:17.4 seconds
	xx:xx:47.4 seconds

After reboot:

1.  slink data after reboot shows 2 timetears:
    a.  -10.5640 second tear at 01:52:10.5640
    b.  +0.5160  second tear at new time 01:51:00.4840
TEST.BP.DP1.40 rate=250 (2011.020 00:00:00.2960 to 2011.020 01:38:35.7640) : 1478867 points, 0.0 msec correction, (min,max,max_step = 0.0,0.0,0.0 msec)
Warning: data overlap: seq=646, expected_time=2011.020 01:52:10.5640, packet_time=2011.020 01:52:00.0000, overlap= -10564.0 msec, input=TEST.BP.DP1.40.D.2011.020.0000.slink
TEST.BP.DP1.40 rate=250 (2011.020 01:47:45.0000 to 2011.020 01:52:00.4840) : 66512 points, -10564.0 msec correction, (min,max,max_step = -10564.0,0.0,10564.0 msec)
TEST.BP.DP1.40 rate=250 (2011.020 01:52:01.0000 to 2011.020 03:00:00.5080) : 1019877 points, 0.0 msec correction, (min,max,max_step = 0.0,0.0,0.0 msec)

If both of these timetears are removed by:
    a.  add a time correction of (-10.5460 + 0.5160) seconds 
	to the data initially timestamped:
	2011.020 01:47:45.0000 to 2011.020 01:52:10.5640
    b.  add a timestamp correction of (0.5160) seconds
	to the following data timestamped:
	2011.020 01:52:00.0000 to 2011.020 01:52:00.4840
it appears based on waveform examination that all of the data are 
present, and that the resulting time tags are correct.

2.  seed data shows no break in the timeseries, BUT:
    a.  it shows a "filled gap" (flat line data) from
	2011.020,00:52:00.0480 and 00:51:01.0000
    b.  based on the ~ 30 second periodic waveform, it appears
	that there is ~ 10 seconds of data missing, eg before
	the 5 second flat line, the transition from 100 Hz to
	1 Hz is at:
	xx:xx:27.4 seconds
	xx:xx:57.4 seconds

3.  slink and seed data synchronized after 2011.020,00:52:01 UTC,
with 100 Hz to 1 Hz transition again at ~
	xx:xx:17.4 seconds
	xx:xx:47.4 seconds




