Database Maintenance and it's effect on replication

Thank you Dynamox.

I apologize - I guess my question then is - does db maintenance (SQL) affect the blocks? ie; would I have to 'block' replicate 50% of my data if the db maintenance 'touches' 50% of the data?

timefinder/srdf only knows blocks ..so which ever blocks change, those are the ones that will be marked to be re-synced.

I guess another question to ask would be - if anyone notices after db maintenance is run, replication takes longer as more blocks need to be synched...

not sure how much will change, depends on database layout, disk layout, what kind of work is done on the database. It is possible that some might notice decrease in performance during that initial sync after the upgrade, after all ..your host and timefinder are reading from the same spindles. If this does not impact your SLAs ..take a look "symqos" feature and maybe setting lower priority on TimeFinder operations for that initial sync after the maintenance work ..it will take longer to sync but will have the least impact on the host IO.

Darrell,
That would really be a question for Microsoft. If they say that the maintenance changes the data at a block level, then yes all of the data changed would need to be replicated.

We have some large SQL databases with nightly SQL maintenance jobs and we see replication (RecoverPoint not SRDF, but also differential block replication) queue up during the maintenance jobs mainly because of the high i/o of the maintenance jobs. The resynch after the jobs usually does not take more than 15 minutes to complete. And on a reindexing of a 2.5TB SQL database that would lead me to believe that at a block level there were not a significant amount of changes.

Aran