K-T Should I be really doing all that every time I make a trivial change to the code? Is there really need to run eupspkg -er decl, setup qserv, stop/start everything? (re the former, it'd be much better if developer could decide, yes, sometimes it is necessary to restart czar or xrootd, but zookeeper or proxy - almost never!) Jacek On 05/30/2014 02:48 PM, Kian-Tat Lim wrote: > Jacek, > > Within the eups framework you have to decide if you want to run > from your build directory or an installed directory. The former is more > convenient (but will probably require changing the Qserv build process a > little bit); the latter is better for provenance and repeatability. > > You also have to decide if you want total safety (rebuild all > configuration and data files on update) or convenience (reuse an > existing instance). I think you always have to stop the existing > instance and start a new one. But it may be possible to have the new > one reuse the configuration and data of the old one. > > What I think you want is this: > > # Assume Qserv configuration/instance in ~/run/qserv > qserv-stop.sh ~/run/qserv > # Assume Qserv source git clone in ~/repo/qserv > cd ~/repo/qserv > eupspkg -er build > eupspkg -er install > eupspkg -er decl -C > setup qserv > qserv-start.sh ~/run/qserv > # and now finally I can run the query that causing troubles > mysql --protocol TCP --port 4040 -A LSST -e "my query" > ######################################################################## Use REPLY-ALL to reply to list To unsubscribe from the QSERV-L list, click the following link: https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=QSERV-L&A=1