0

Trying to figure out if there's way to replicate a subset of tables (columns) from HADR Standby, with ROS enabled. Latency of O(10) sec can be tolerated. We are using luw V10.5FP8 right now and will upgrade to V11 at some point.

I understand the read-only limitation on HADR Standby, and that eliminates some options, eg. QRepl / infosphere CDC, which write monitoring/metadata back to source.

Furthermore, assuming we still limit ourselves to db2 heterogeneous env and the repl user have read access to all the tables/columns, is there a replication tool that doesn't depends on constant source db connection? Meaning such tool only scans transaction log files, and writes to its own external metadata/file, without bothering with source db at all? It would be even better if it can capture once, and replay in multiple targets.

Really appreciate your inputs.

fall14123
  • 61
  • 1
  • 8
  • Is there a good reason for not doing the replication from the primary, or to avoid introducing a second standby ? There seems no obvious easy solution for doing partial replication from an ROS standby, apart from custom scripting. – mao Oct 18 '17 at 07:02
  • Thanks for your replay @mao. Greg answered my question in http://www.idug.org/p/fo/st/topic=19&post=183347#p183347 as well. The Standby restriction on db2ReadLog API does make it difficult. The motivation was to distribute QCapture load around, since are a little conservative about running multiple QCapture program on our OLTP source, mostly due to performance concerns. We don't use shared storage for QRepl at the moment, which means all the Qs are collocated with the OLTP source. – fall14123 Oct 18 '17 at 21:55
  • and if you don't mind, could you please elaborate a little bit more on "apart from custom scripting" part? thanks again :) – fall14123 Oct 18 '17 at 21:57
  • If you mention a log reader, there is the API db2ReadLogNoConn (i.e. no database connection). The docs don't mention if that will run on a standby . Regardless, developing c/c++ code may not be cost effective. If the business-requirement for the subset of tables/columns does not require near-real-time accuracy (i.e. can accept 30minute or hourly refresh) then a simple script to pull from standby and push to new target just might suffice (depending on data volumes and latency etc). Cheap and cheerless, but possibly tactical. – mao Oct 19 '17 at 07:04
  • Nod....... i'm not familiar with c/c++ at all ...... wondering if there are already products offering such feature, meaning doing ETL/replication using only log files without bothering the source db at all....... – fall14123 Oct 19 '17 at 15:01

0 Answers0