This issue goes hand in hand with
http://code.google.com/p/gbif-indexingtoolkit/issues/detail?id=161
The Data Portal logs some common user activity inside the portal. This is
written currently to the
live DB, as seen on the portal.properties file. All user logs are written into
the "gbif_log_message"
table.
As a first step, we need to extract these logs to another DB (that could share
the same DB
structure as the one in the live DB). This will essentially make the live DB
almost read-only (for
small exceptions) and we can accomplish separation of concerns between logging
and primary
biodiversity data.
Would a change in the portal.properties
logDataSource.username=username
logDataSource.password=password
logDataSource.url=jdbc:mysql://LIVE_DB:3306/portal?
autoReconnect=true&useUnicode=true&characterEncoding=UTF8&characterSetResults=UT
F8
be sufficient? This needs to be tested first.
As a second step, the HIT is also going to write to this new DB all the
indexing logs. So in the
portal side, there should be special care taken at the moment of reading the
"indexing history"
and "user logs" so that this requests get directed to the new DB and not to the
live or indexing
DB.
Examples that need special care:
http://data.gbif.org/datasets/provider/308/logs/
http://data.gbif.org/datasets/provider/308/indexing/
http://data.gbif.org/datasets/provider/4/logs/?resource=&event=3001-
3999&logGroup=&logLevel=&sd_day=27&sd_month=05&sd_year=2009&ed_day=27&ed_month
=05&ed_year=2010
etc...
Would a change in portal.properties
harvestingDataSource.username=username
harvestingDataSource.password=password
harvestingDataSource.url=jdbc:mysql://INDEXING_DB:3306/portal?
autoReconnect=true&useUnicode=true&characterEncoding=UTF8&characterSetResults=UT
F8
also be sufficient? Needs to be tested...
Original issue reported on code.google.com by josecua...@gmail.com on 27 May 2010 at 2:30
Original issue reported on code.google.com by
josecua...@gmail.com
on 27 May 2010 at 2:30