Reduntant zip_size handling at delta application time
| Affects | Status | Importance | Assigned to | Milestone | |
|---|---|---|---|---|---|
| Percona XtraBackup moved to https://jira.percona.com/projects/PXB |
Fix Released
|
High
|
Laurynas Biveinis | ||
| 2.0 |
Fix Released
|
High
|
Laurynas Biveinis | ||
| 2.1 |
Fix Released
|
High
|
Laurynas Biveinis | ||
Bug Description
There is no need for a correct fil_space_create() flags value: it is never queried in the short-lived XtraBackup InnoDB file system instance, that is live only around delta application time. A valid flag value avoids assertion failures inside fil_space_create(), but it simply enough for it to be any correct value (e.g. 0).
This is alternate (better) fix for bug 1028949, regression from bug 932623, will also fix bug 1044398.
Note that zip_size cannot be fully removed from incremental backup tablespace create because fix for bug 1043762 still requires it for the tablespace creation.
Thus the correct handling would be:
1) zip_size is written to the incremental backup tablespace metadata;
2) zip_size is optional during the metadata read;
3) fil_space_create is always passed 0 for tablespace flag as it's used only for file name hashing;
4) the actual tablespace create computes the tablespace flags from zip_size if it was read.
Related branches
- Alexey Kopytov (community): Approve
-
Diff: 304 lines (+139/-56)1 file modifiedsrc/xtrabackup.c (+139/-56)
- Alexey Kopytov (community): Approve
-
Diff: 370 lines (+181/-61)5 files modifiedsrc/innodb_int.c (+116/-0)
src/innodb_int.h (+35/-1)
src/write_filt.c (+2/-17)
src/xtrabackup.c (+25/-41)
src/xtrabackup.h (+3/-2)
| description: | updated |
| description: | updated |
| Shahriyar Rzayev (rzayev-sehriyar) wrote | #1 |
Percona now uses JIRA for bug reports so this bug report is migrated to: https:/