Custom Query (541 matches)


Show under each result:

Results (31 - 33 of 541)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Ticket Resolution Summary Owner Reporter
#493 fixed ewmseedarchiver path on Windows et stefan

I added to the .d file the following for now, but fixing this at some point would be nice:

#Archive /data/archive/%n/%s/%Y.%j.%n.%s.%l.%c.mseed
# The Windows compile doesn't like forward slashes nor drive letters; 
# something like this worked. (note double, triple and single slashes!)
# Only one directory-level substitution appears to work on Windows.
# Archive  ..\\data\\msarchive\\\%s\%Y.%j.%n.%s.%l.%c.mseed

#501 fixed tbuf2mseed silently fails et stefan

tbuf2mseed can not be used to create continuous archives, it fails after a few days.

At around 13:23 tbuf2mseed silently failed. (At least that's the final time stamp on the output of ewmseedarchiver.) See

The module stays alive and outputs to the tbuf2mseed log but ewmseedarchiver doesn't produce any mseed until tbuf2mseed is restarted.

This is on a Windows Server 2008 R2 system.

#509 fixed tbuf2mseed 'Unable to represent difference in <= 30 bits' errors et et

When the 'tbuf2mseed' module is receiving data from station UW_NNVM, errors like these are generated out of the 'libmseed' library:

20150223_UTC_23:50:12 Debug: NNVM.HNN.UW.--  ns=200  dt=200.0  type=i4
Error: msr_pack_steim2(UW_NNVM__HNN_D): Unable to represent difference in <= 30 bits
Error: msr_pack(UW_NNVM__HNN_D): Error packing record
20150223_UTC_23:50:12 Cannot pack trace buffer or send records!

This issue is somewhat related to #501 ("tbuf2mseed silently fails"), as the errors seem tied to memory leaks in the module.

This fix is proposed by Chad:

The "Unable to represent difference in <= 30 bits" error is libmseed complaining that it cannot create miniSEED from the data with Steim2 compression. Likely the data contain spikes or other full scale signals that cannot be represented in Steim2. If this is happening a lot memory will be leaked and eventually become a big problem as Larry described. You can retain the signal in the data (and avoid the gaps being generated by the errors) if you switch to Steim1 compression, which can represent a full 32-bit sample difference. This will also avoid the memory leak by avoiding that particular error. Unfortunately, tbuf2mseed/ringdup.c appears to be hardcoded to use Steim2 for integer samples. The effected part of the code is similar to that in ew2ringserver; if you look at ew2ringserver.c you can see how I've extended it to allow the user to specify the encoding for 32-bit integers. From the operator perspective there is a Int32Encoding parameter in the .d file that can be STEIM2, STEIM1 or INT32 (no compression).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Note: See TracQuery for help on using queries.