* all metadata in Zookeeper

* fast, transient cache for czar, implemented through
   in-memory mysql table, the table kept as part of the
   mysql database that is sitting next to czar

* the cache synchronized with Zookeeper by
   a dedicated python-based daemon (when data changed)

* c++ qserv czar code fetches metadata from
   the transient cache (querying mysql)

* some tbd synchronization to avoid updating data
   when czar is reading (I have a couple ideas, but
   that is too much typing, I can explain tomorrow).

* no c/c++ zookeeper client needed

Thoughts? Too complicated?

Jacek

########################################################################
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