kernel_mutex optimization around trx_create() and trx_free()
Bug #747112 reported by
Yasufumi Kinoshita
This bug affects 1 person
| Affects | Status | Importance | Assigned to | Milestone | ||
|---|---|---|---|---|---|---|
| MySQL Server |
Unknown
|
Unknown
|
||||
| Percona Server moved to https://jira.percona.com/projects/PS | Status tracked in 5.7 | |||||
| 5.1 |
Won't Fix
|
Wishlist
|
Unassigned | |||
| 5.5 |
Won't Fix
|
Wishlist
|
Unassigned | |||
| 5.6 |
Fix Released
|
Wishlist
|
Unassigned | |||
| 5.7 |
Invalid
|
Undecided
|
Unassigned | |||
Bug Description
If trx_create() and trx_free() don't need kernel_mutex, they should be unprotected by the kernel_mutex.
The fix should be effective for the workload that
"small" and "read-only" transaction "storm" (much parallelized, many).
http://
(Though also mentioned about readview and kernel_mutex in this report,
I think even only trx_create and trx_free are enough effective for the storm for now)
| Changed in percona-server: | |
| status: | New → Confirmed |
| Changed in percona-server: | |
| importance: | Undecided → Wishlist |
| status: | Confirmed → Triaged |
| tags: | added: innodb performance upstream |
Revision history for this message
| Shahriyar Rzayev (rzayev-sehriyar) wrote | #1 |
To post a comment you must log in.
Percona now uses JIRA for bug reports so this bug report is migrated to: https:/