On Tue, Jun 25, 2019 at 2:19 AM Prabhat Sahu <prabhat.sahu@enterprisedb.com> wrote: > I have tested the TOAST patches(v3) with different storage options like(MAIN, EXTERNAL, EXTENDED, etc.), and > combinations of compression and out-of-line storage options. > I have used a few dummy tables with various tuple count say 10k, 20k, 40k, etc. with different column lengths. > Used manual CHECKPOINT option with (checkpoint_timeout = 1d, max_wal_size = 10GB) before the test to avoid performance fluctuations, > and calculated the results as a median value of a few consecutive test executions.
Thanks for testing.
> All the observation looks good to me, > except for the "Test1" for SCC UPDATE with tuple count(10K/20K), for SCC INSERT with tuple count(40K) there was a slightly increse in time taken > incase of "with patch" result. For a better observation, I also have ran the same "Test 1" for higher tuple count(i.e. 80K), and it also looks fine.
Did you run each test just once? How stable are the results?
No, I have executed the test multiple times(7times each) and calculated the result as the median among those,