HBASE-29479: QuotaCache should always return accurate information #7188
Conversation
This file contains hidden or 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 comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
charlesconnell
commented
Aug 27, 2025
Comment on lines
-84
to
-87
| // No write throttling | ||
| configureLenientThrottle(ThrottleType.ATOMIC_WRITE_SIZE); | ||
| refreshQuotas(); | ||
|
|
There was a problem hiding this comment.
These lines shouldn't have been here. Seting a per-user quota of any type shadows the default quotas of all types, making the following assertion fail, if QuotaCache is working properly. The updates to QuotaCache in this PR exposed this test as incorrect.
| private QuotaRefresherChore refreshChore; | ||
| private boolean stopped = true; | ||
|
|
||
| private final Fetcher<String, UserQuotaState> userQuotaStateFetcher = new Fetcher<>() { |
There was a problem hiding this comment.
These Fetcher objects have the same logic as before, but are lifted into top-level-class fields so I can access them in new places.
Comment on lines
+204
to
+210
| String user = getQuotaUserName(ugi); | ||
| if (!userQuotaCache.containsKey(user)) { | ||
| userQuotaCache.put(user, | ||
| QuotaUtil.buildDefaultUserQuotaState(rsServices.getConfiguration(), 0L)); | ||
| fetch("user", userQuotaCache, userQuotaStateFetcher); | ||
| } | ||
| return userQuotaCache.get(user); |
There was a problem hiding this comment.
This is the meat of the change. If a caller asks for a quota for a given user, check if the quota is already saved in the cache. If so, return it from the cache. If not, do a lookup from the hbase:quota table, cache the result, and then return it. This means that users of QuotaCache don't need to wait for its chore to run in order to get correct results.
| } | ||
|
|
||
| static interface Fetcher<Key, Value> { | ||
| Get makeGet(Map.Entry<Key, Value> entry); |
There was a problem hiding this comment.
None of the Fetcher implementations used the entry value, so I've stopped passing the a fully Entry
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
🎊 +1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
rmdmattingly
approved these changes
Sep 2, 2025
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 asked for a QuotaLimiter for a given user, the QuotaCache checks if that user's quota state is cached. If it is, it returns a QuotaLimiter based on the cached information. If not, it inserts default quota state into the cache, and returns a default QuotaLimiter. On the next run of the QuotaRefresherChore, it looks in the hbase:quota table for the quota settings for every user in the cache. The consequence of this is: from when the QuotaCache is first asked about a user, until the next run of QuotaRefresherChore, the QuotaLimiters for that user are wrong.
I think it's unacceptable for QuotaCache to return incorrect information. I can see from a code comment that it was implemented this was as a performance optimization, to avoid doing a Get on the quota table in the query handler path. However, the performance consequences of not enforcing quotas can be a lot worse that a few Gets.
In this ticket I'm adding logic to QuotaCache so it immediately fetches quota information for a user from the quota table any time that user is not in its cache.