Optimise the restoration process #57
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
shreyas-s-rao
requested review from
georgekuruvillak and
swapnilgm
as code owners
|
|
swapnilgm
added
kind/enhancement
shreyas-s-rao
changed the title
|
As discussed last week, can we include some tests please? ;-) |
georgekuruvillak
suggested changes
Oct 11, 2018
swapnilgm
added
the
reviewed/do-not-merge
amshuman-kr
reviewed
Oct 12, 2018
shreyas-s-rao
force-pushed
the
restore-parallelization
branch
from
7126058
to
c92de4c
Compare
|
Code changes are finished and ready for review. Will push the unit tests shortly. |
swapnilgm
suggested changes
Oct 26, 2018
shreyas-s-rao
force-pushed
the
restore-parallelization
branch
from
f8a7699
to
8c4de61
Compare
|
Will push unit tests shortly. |
shreyas-s-rao
changed the title
amshuman-kr
reviewed
Nov 12, 2018
| errCh <- fmt.Errorf("snap index mismatch for delta snapshot %d; expected snap index to be atleast %d", fetchedSnapIndex, nextSnapIndexToApply) | ||
| return | ||
| } | ||
| if fetchedSnapIndex == nextSnapIndexToApply { |
There was a problem hiding this comment.
What if fetchedSnapIndex > nextSnapIndexToApply?
There was a problem hiding this comment.
It will just continue to the next iteration. The logic is: when fetchedSnapIndex == nextSnapIndexToApply, it applies that snap as well as any existing subsequent snaps till an unfetched snap is encountered, and it moves to the next iteration. This way, all snaps are applied in the right order.
| DeltaSnapList: deltaSnapList, | ||
| } | ||
| err = rstr.Restore(restoreOptions) | ||
| Expect(err).Should(HaveOccurred()) |
There was a problem hiding this comment.
Might be good to validate the error message also.
There was a problem hiding this comment.
Sure, will definitely look into it.
swapnilgm
removed
needs/rebase
|
LGTM |
shreyas-s-rao
force-pushed
the
restore-parallelization
branch
from
e72ce5c
to
2d1d7b7
Compare
swapnilgm
reviewed
Nov 13, 2018
pkg/snapshot/restorer/restorer.go
Outdated
| r.logger.Infof("Cleanup complete") | ||
| } | ||
|
|
||
| // fetch fetches delta snapshots as events and persists them onto disk. |
pkg/snapshot/restorer/restorer.go
Outdated
|
|
||
| // fetch fetches delta snapshots as events and persists them onto disk. | ||
| func (r *Restorer) fetchSnaps(fetcherIndex int, fetcherInfoCh <-chan fetcherInfo, applierInfoCh chan<- applierInfo, snapLocationsCh chan<- string, errCh chan<- error, stopCh chan bool, wg *sync.WaitGroup) { | ||
| defer func() { |
There was a problem hiding this comment.
Refactor:
defer wg.Done()
pkg/snapshot/restorer/restorer.go
Outdated
| } | ||
| } | ||
|
|
||
| // apply applies delta snapshot events to the embedded etcd sequentially, in the right order of snapshots, regardless of the order in which they were fetched. |
pkg/snapshot/restorer/restorer.go
Outdated
|
|
||
| // apply applies delta snapshot events to the embedded etcd sequentially, in the right order of snapshots, regardless of the order in which they were fetched. | ||
| func (r *Restorer) applySnaps(client *clientv3.Client, remainingSnaps snapstore.SnapList, applierInfoCh <-chan applierInfo, errCh chan<- error, stopCh <-chan bool, wg *sync.WaitGroup) { | ||
| defer func() { |
pkg/snapshot/restorer/restorer.go
Outdated
|
|
||
| wg.Add(1) | ||
|
|
||
| eventsList := make([][]event, len(remainingSnaps)) |
There was a problem hiding this comment.
Do we really need eventsList? It is kind of invalidating purpose of persisting fetched delta snapshot locally, isn't it? This way at one stage we might have data from all delta snapshots in memory, which will lead to unnecessary memory consumptions and OOM issues. Please read the persisted snapshots in memory in order of index and apply it. If for current index, snapshot is not yet fetched, wait until it gets fetched.
swapnilgm
added
the
reviewed/ok-to-test
gardener-robot-ci-1
added
needs/ok-to-test
swapnilgm
added
reviewed/ok-to-test
swapnilgm
approved these changes
Nov 14, 2018
gardener-robot-ci-1
added
needs/ok-to-test
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.
What this PR does / why we need it:
This PR optimises the restoration process by parallelizing the fetching of delta snapshots.
Which issue(s) this PR fixes:
Fixes #41
Special notes for your reviewer:
Release note: