[evlatests] VisPipe/IDCAF testing
Pete Whiteis
pwhiteis at nrao.edu
Fri Jul 6 19:30:25 EDT 2007
Several hours were spent today testing IDCAF->Vispipe processor
interactions.
Briefly:
- During continuum mode, life is good for both IDCAF and Vispipe.
Startups and shutdowns from both ends seem to work quickly and reliably.
- Spectral line modes require a little more time for things to get
organized when a client disconnects/reconnects to Vispipe. Sometimes
this can take up to 2.5 minutes. The reason for this is that the Array
Processor device driver interface shuts down and has to restart. The
results in both the driver resyncing to the data stream and the user
application resyncing to the driver. For the time being, this means
that client connecting in to vispipe will need to exercise patience.
In the longer term, some re-organization of the vispipe processor
application may help to reduce re-sync time.
- Sometimes mode changes result in unusual behavior, such as an
inexplicable change in data transfer periods, or data size. This can
cause client software frustration, but nothing more severe.
- For system debug black-belts: 1 visapp process means no-one has
connected to the vispipe app. 2 visapp processes indicates that a
client task was formerly connected to vispipe app, but has since
disconnected. 3 visapp processes indicates that a client has
successfully connected and is either receiving data or waiting for
vispipe app to re-sync.
- Walter found and fixed a couple instances of undesirable behavior in
IDCAF. I'll let him speak to this.
-Pete-
-------------- next part --------------
A non-text attachment was scrubbed...
Name: pwhiteis.vcf
Type: text/x-vcard
Size: 144 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20070706/4f10b1df/attachment.vcf>
More information about the evlatests
mailing list