Make DB parameters for treeherder-dev similar to treeherder-stage
Categories
(Tree Management :: Treeherder: Infrastructure, task, P3)
Tracking
(Not tracked)
People
(Reporter: armenzg, Unassigned)
References
Details
Currently these are the innodb_ differences between the three:
| Parameter | Production | Stage | Dev |
|---|---|---|---|
| innodb_io_capacity_max | 5000 | 5000 | 2500 |
| innodb_lru_scan_depth | 256 | 256 | 256 |
| innodb_log_file_size | 17179869184 (16GB) | 17179869184 (16GB) | 8589934592 (8GB) |
| innodb_read_io_threads | 8 | 8 | 4 |
| innodb_write_io_threads | 8 | 8 | 4 |
| innodb_buffer_pool_size | 48318382080 (45GB) | 11811160064 (11GB) | 11811160064 (11GB) |
We need to bring treeherder-dev closer to treeherder-stage since they have the same DB size, storage type & EC2 instance type.
Specifically, I want to change innodb_io_capacity_max, innodb_log_file_size, innodb_read_io_threads & innodb_write_io_threads.
This is assuming that New Relic shows that treeherder-stage is better performant.
Parameter|Production|Stage/Dev
-----------|-------------|------|----
innodb_io_capacity_max|5000|same
innodb_lru_scan_depth|256|same
innodb_log_file_size|17179869184 (16GB)|same
innodb_read_io_threads|8|same
innodb_write_io_threads|8|same
innodb_buffer_pool_size|48318382080 (45GB)|11811160064 (11GB)
bobm, what are your thoughts?
If you need to I have added a comment in here and there are 3 attachments with the innodb_ values for each.
| Reporter | ||
Updated•6 years ago
|
| Reporter | ||
Updated•6 years ago
|
Updated•10 days ago
|
Description
•