Estimated pg_dump and pg_restore Times for Aurora PostgreSQL

I measured the time to pg_dump and pg_restore Aurora PostgreSQL data to/from a bastion server on EC2. Note that network bandwidth and disk I/O performance vary by instance type, disk size, and IOPS settings, so treat these as rough estimates.

Results

No	Test Pattern	DB Name	DB Size (GB)	Data Content	Start Time	End Time	Duration
1	pg_dump	postgres	729GB	Aozora Bunko text data	14:31:33	19:01:44	4:30:11
2	(parallelism=1)	tpch	45GB	HammerDB TPC-H data	12:59:44	13:34:22	0:34:38
3		tpcc	118GB	HammerDB TPC-C data	13:43:07	14:26:32	0:43:25
4		blob	98GB	Binary in bytea (25MB file × 4000 files)	9:26:14	12:38:51	3:12:37
5	pg_dump	postgres	729GB	Same as above	18:30:26	19:59:33	1:29:07
6	(parallelism=8)	tpch	45GB	Same as above	18:11:08	18:28:35	0:17:27
7		tpcc	118GB	Same as above	13:16:13	13:44:05	0:27:52
8		blob	98GB	Same as above	20:01:51	21:04:38	1:02:47
9	pg_restore	postgres	729GB	Same as above	21:09:34	0:26:25	3:16:51
10	(parallelism=8)	tpch	45GB	Same as above	20:58:25	21:06:24	0:07:59
11		tpcc	118GB	Same as above	20:46:58	20:56:44	0:09:46
12		blob	98GB	Same as above	12:45:23	13:07:43	0:22:20

pg_dump Command with Parallelism 1

pg_dump -h aurorapgsqlv1.cluster-xxxx.ap-northeast-1.rds.amazonaws.com -U postgres -Fc postgres > postgres.dump

pg_dump Command with Parallelism 8

pg_dump -h aurorapgsqlv1.cluster-xxxxx.ap-northeast-1.rds.amazonaws.com -j 8 -U postgres -F d -f /data/postgres postgres

pg_restore Command with Parallelism 8

pg_restore -h aurorapgsqlv1.cluster-xxxxx.ap-northeast-1.rds.amazonaws.com -U postgres -j 8 -d postgres postgres.dump