Hello,
Is someone here who implements a concurrent gather strategy for stats
collection for a big database that happens to be in an active
dataguard configuration?
My issue is: during concurrent stats execution in the primary database
( usually during a 6 hour window) the standby starts to lag behind almost
3 hours (no other major activity during this time in both primary or
standby). I`m curios if someone faced this issue before.
Some technical details about the issue:
-> on primary DB, on node where the concurrent stats run 25% of total wait
time is spent into "library cache lock"
-> on standby PR00 process wait mainly for "latch free" , following the
chain is blocked by LGWR who waits for "enq: IV contention"
->top latch in standby is held by LGWR with the name "readable standby
metadata redo cache"
Thank you.