Fix initial sync race #208
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
lkysow
force-pushed
the
sync-catalog-tests
branch
2 times, most recently
from
bb115b6 to
8be4a97
Compare
lkysow
force-pushed
the
fix-init-race
branch
3 times, most recently
from
660aaa6 to
29ec319
Compare
|
thks @lkysow looks good to me |
lkysow
force-pushed
the
fix-init-race
branch
3 times, most recently
from
7c43b7e to
27472e2
Compare
adilyse
approved these changes
Mar 6, 2020
catalog/to-consul/syncer_test.go
Outdated
Comment on lines
173
to
174
| // Create a service directly in Consul. Since it was created directly we | ||
| // expect it to be deleted but not until the initial sync is performed. |
There was a problem hiding this comment.
Suggested change
| // Create a service directly in Consul. Since it was created directly we | |
| // expect it to be deleted but not until the initial sync is performed. | |
| // Create a service directly in Consul. Since it was created on the | |
| // synthetic sync node and has the sync-associated tag, we expect | |
| // it to be deleted but not until the initial sync is performed. |
A service needs the sync tag and to be registered on the sync node or the sync process would ignore it.
catalog/to-consul/syncer_test.go
Outdated
| require.NoError(t, err) | ||
|
|
||
| // We wait until the syncer has had the time to delete the service. | ||
| time.Sleep(200 * time.Millisecond) |
There was a problem hiding this comment.
It seems like we might run into edge cases having this value exactly equal the SyncPeriod in the test ConsulSyncer. What about setting up the ConsulSyncer with an even smaller interval?
There was a problem hiding this comment.
Yes! Great idea. I didn't like this sleep but a retry loop doesn't work because it might run through before syncFull has run :D
A race condition was occurring where we would delete services via the `watchReapableServices` function if `Sync` wasn't called fast enough by `resource.go`. This occurred because `watchReapableServices` makes a call to the catalog and then deletes everything with a k8s tag that isn't in its `s.services` map. If `Sync` isn't called, then this map is empty and so it tries to delete everything. I've solved this by blocking until `Sync` is first called.
adilyse
reviewed
Mar 6, 2020
| Address: a.HTTPAddr, | ||
| }) | ||
| require.NoError(t, err) | ||
| s, closer := testConsulSyncer(client) |
There was a problem hiding this comment.
Suggestion:
s, closer := testConsulSyncerWithConfig(client, func(s *ConsulSyncer) {
s.SyncPeriod = 5 * time.Millisecond
})
Since the testConsulSyncer also Runs the syncer, the sync period change may not get picked up if it's made afterwards.
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.
A race condition was occurring where we would delete services via the
watchReapableServicesfunction ifSyncwasn't called fast enough byresource.go. This occurred becausewatchReapableServicesmakes acall to the catalog and then deletes everything with a k8s tag that
isn't in its
s.servicesmap. IfSyncisn't called, then this mapis empty and so it tries to delete everything.
I've solved this by blocking until
Syncis first called.Closes #191