Date
Attendees
Goals
- Get together, resolve bottlenecks, make progress faster, plan, design, exchange ideas
Discussion items
Add "data version" in zookeeper, to ensure the code we are executing is compatible with data in zookeeper
Related to DM-1338
- Use case:
- I'm working with qserv branch x, and I have run directory for that branch
- I'm now creating qserv branch y
- generating new run directory is slow, I want to reuse data from run directory that I created for branch x
- Currently, we regenerate everything (that is "qserv_configure --all")
- Add flag "–keepdata" - it will regenerate everything but keep data (mysql data, zookeeper data), including scisql plugin
- Note that this means we will generate config files (with correct paths) bases on eups-managed environment, but we will not depend on how eups is configured when we are starting services. It is more robust that way
qserv-char.sh
- qserv-czar.sh currently creates qservw_worker
- that works for mono-node, but will break for multi-node setup
- need to fix, need qserv-worker.sh
DM-1251
- where to keep the data about qserv metadata?
- in one dedicated mysql instance. InnoDB (or even in-memory engine), master-slave replication
DM-318
- some testing done, no new code, we are ok with what we have now. Nothing to review. Can close
DM-1210
- need to remerge/retest because master changed
- then DM-1249
DM-631
- will be done end of this week, review should be fast
DM-215
- not done, but started, move to Nov
DM-1445
- check fi Qserv compiles with C++11
- probably we didn't see any issues because we don't depend on sconsUtils
other todos
- get jira account for Vaikunth
- plan Fabrice's work for Nov
- revisit stories in DM-1030 (refactoring)