Timeline for How to improve insert in RDS of AWS?
Current License: CC BY-SA 3.0
7 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 27, 2017 at 18:52 | comment | added | Rick James | Sounds like you have a good test case for comparing timings on 100*10000 vs 1000*1000 vs 10000*100. | |
| Jan 27, 2017 at 10:19 | comment | added | Ondrej Burkert | For the record I went with 10k and it went just fine yesterday. But wanted to know more what I was doing:) | |
| Jan 27, 2017 at 10:18 | comment | added | Ondrej Burkert | Thanks for answering:). In my case the VM is not so tiny so I was just wondering what to base the batch size on. Making one million inserts in batches of 100 wouldn't be practical. | |
| Jan 26, 2017 at 21:59 | comment | added | Rick James | And your tiny VM is even more likely to hit some limit. Play it safe: 100 rows/batch (90% of theoretical max). | |
| Jan 26, 2017 at 21:57 | comment | added | Rick James | (1) diminishing returns -- 1K: 99.0% of theoretical max versus 10K: 99.9%. (2) 10K may hit other limits (log file size, max packet, etc) that would (a) slow it down, maybe by a factor of 2, or (b) cause to fail. | |
| Jan 26, 2017 at 7:54 | comment | added | Ondrej Burkert | May I ask why 100-1000 is a good range for "batch" inserts? I was just about to try 10000 in a batch... | |
| Oct 19, 2016 at 1:09 | history | answered | Rick James | CC BY-SA 3.0 |