Apologies to everyone - my numbers were wrong! I used a script to pull them and I didn't look closely enough at the code to see that it was doing something different than I expected. Rich's script wouldn't account for some logs which have multiple entries in v$archived_log - but it shouldn't throw off the average too much, so I ran it for comparison sake with other numbers in this thread. 30 agents / 500 targets :: 265 MB/hr 500 agents / 2000 targets :: 1.7 GB/hr 350 agents / 2000 targets :: 2 GB/hr FYI, someone emailed me off list with these numbers: 1000 agents / 18000 targets :: 5-8 GB/hr Finally, summarizing other data from this thread: 5 agents / 86 targets :: 120 MB/hr [12.1.0.3] 11 agents / 72+ targets :: 137 MB/hr 40 agents / 750 targets :: 2GB/hr [Max - is this correct? Did you use the redo block size or db block size?] All numbers are from 12.1.0.4 unless otherwise noted. Anecdotally, I've also heard that old versions of agents can skew numbers a bit -- I know that there a few old agents involved in my numbers. -- http://about.me/jeremy_schneider On Wed, Nov 12, 2014 at 9:18 AM, Wenping Bo <wbo1@xxxxxxxxxxxxxx> wrote: > For comparison purpose, our 12.1.0.4 EM repository redo generates about > 137MB/hr. > > - 7 DB instances > - 11 hosts /6 listeners / 11 agents > - 72 targets with status > > Thanks Rich for the nice query. > > Wenping > > > > ----- Original Message ----- > From: "Rich Jesse" <rjoralist3@xxxxxxxxxxxxxxxxxxxxx> > To: oracle-l@xxxxxxxxxxxxx > Sent: Tuesday, November 11, 2014 4:03:54 PM > Subject: Re: OEM repository redo generation > > Jeremy posted: > >> Quick question for someone out there with an OEM repository - how many >> targets and hosts do you have, and what's your typical hourly redo >> generation rate on the repository database? > > For a smaller scale comparison, my 12.1.0.3 EM does 120MB/hr redo with: > > - 18 DB instances > - 5 hosts / 5 listeners / 5 agents > - 86 total targets / 64 targets with status > > I used this to determine average redo generation on the EM12c repo DB: > > SELECT > (SUM(BLOCKS*BLOCK_SIZE)/1024/1024 / (MAX(NEXT_TIME) - MIN(FIRST_TIME))) / 24 > REDO_MB_PER_HOUR > FROM V$ARCHIVED_LOG; > > Rich -- //www.freelists.org/webpage/oracle-l