creating a new datastore

dsprodtool --add IPP-MOPS-TEST --type IPP-MOPS --description 'not described here'

telling publish about the datastore

pubtool -defineclient -stage diff -product IPP-MOPS-TEST -workdir neb://any/gpc1/IPP-MOPS-TEST -unmagicked

you may have to use pubtool to turn on the client once it is created

setting up a password on the datastore

in /data/ippdb02.0/datastore/dsroot/IPP-MOPS-TEST

the .htaccess file is:

AuthType Basic
AuthName "Restricted Files"
AuthBasicProvider file
AuthUserFile /data/ippdb02.0/datastore/dsroot/IPP-MOPS-TEST/config/password
Require user mops

password file is created using htpasswd

htpasswd -c password mops

reprocessing the data

Queue the chips:

chiptool -dbname gpc1 -definebyquery -set_end_stage warp -set_tess_id RINGS.V0 -obs_mode 3Pi -set_dist_group ThreePi.MopsTest.20101018 -dateobs_begin 2010-10-11T00:00:00 -dateobs_end 2010-10-13T00:00:00 -set_reduction MOPS.TEST -obs_mode %3Pi% -simple -set_label MopsTest.20100708 -set_workdir !neb://@HOST@.0/gpc1/ThreePi.MopsTest.20100708

(but be smarter about the label/directory names)

Queue the diffs using survey.add.diff or survey.add.WSdiff (depending on what is needed. Note that for OSS, care must be taken to select the proper warps to diff.

Using the publishing pantasks

• Currently, the publishing.trigger tasks queues every label in pantasks for all of the active publishing clients. This means if you created IPP-MOPS-TEST for the sole purpose of publishing MopsTest.20100708, you will need to remove the other labels or else they will also be queued up to be published under IPP-MOPS-TEST.

• There was a bug: if you take out all the labels in pantasks, publishing.trigger will queue up everything, regardless of label. I have fixed this, and it only queues if there are Labels present.

• publishing.load does not respect labels: anything queued in PublishRun and with an active client_id is processed, no matter what the label.