importccl: speed up revert of IMPORT INTO empty table #52754
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
This change is |
pbardea
approved these changes
Aug 13, 2020
pkg/ccl/importccl/import_stmt.go
Outdated
| details.Walltime = p.ExecCfg().Clock.Now().WallTime | ||
|
|
||
| // Check if the tables being imported into are starting empty, in which | ||
| // case we can a cheap clear-range instead of revert-range to cleanup. |
There was a problem hiding this comment.
nit: "... we can a cheap ..."
pkg/ccl/importccl/import_stmt.go
Outdated
| for i := range details.Tables { | ||
| if !details.Tables[i].IsNew { | ||
| tblSpan := sqlbase.NewImmutableTableDescriptor(*details.Tables[i].Desc).TableSpan(keys.TODOSQLCodec) | ||
| res, err := p.ExecCfg().DB.Scan(ctx, tblSpan.Key, tblSpan.EndKey, 1) |
dt
force-pushed
the
import-into-empty
branch
6 times, most recently
from
c3d663e
to
c4333ee
Compare
When IMPORT INTO fails, it reverts the tables to their pre-IMPORT state. Typically this requires running a somewhat expensive RevertRange operation that finds the keys written by the IMPORT in amongst all the table data and deletes just those keys. This is somewhat expensive -- we need to iterate the keys in the target table and check them to see if they need to be reverted. Non-INTO style IMPORTs create the table into which they will IMPORT and thus can just drop it wholesale on failure, instead of doing this expensive revert. However INTO-style IMPORTs could use a similarly fast/cheap wholesale delete *if they knew the table was empty* when the IMPORT was started. This change tracks which tables were empty when the IMPORT started and then deletes, rather than reverts, the table span on failure. Release note (performance improvement): Cleaning up after a failure during IMPORT INTO a table which was empty is now faster.
|
bors r+ |
|
Build succeeded: |
This was referenced Mar 30, 2021
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When IMPORT INTO fails, it reverts the tables to their pre-IMPORT state.
Typically this requires running a somewhat expensive RevertRange operation
that finds the keys written by the IMPORT in amongst all the table data
and deletes just those keys. This is somewhat expensive -- we need to
iterate the keys in the target table and check them to see if they
need to be reverted.
Non-INTO style IMPORTs create the table into which they will IMPORT and
thus can just drop it wholesale on failure, instead of doing this expensive
revert. However INTO-style IMPORTs could use a similarly fast/cheap
wholesale delete if they knew the table was empty when the IMPORT was
started.
This change tracks which tables were empty when the IMPORT started and
then deletes, rather than reverts, the table span on failure.
Release note (performance improvement): Cleaning up after a failure during IMPORT INTO a table which was empty is now faster.