Oracle DB and Takeover Impact
Hello all,
We are running Oracle DB 11.2.0.4 (on RAC One-Node, Solaris 11.2 SPARC) connected to clustered ZFS 7320's.
Protocol is dNFS (v3). Single pool, 16 shares.
We have previously tested storage head take-overs successfully (with database running) but testing under real-world load was not possible. We are now planning to upgrade the ZFS software which is out of date (2013.06.05), and cluster is serving production traffic. Is there a best practice or steps to follow that might help alleviate negative impact on the databases during takeover and fail-back (other than obviously doing it off-peak)? We are, of course, hoping to avoid a DB shutdown if possible.