HDDS-11764. Fix initiating DocumentBuilderFactory performance problem. #7462
Conversation
adoroszlai
left a comment
adoroszlai
left a comment
There was a problem hiding this comment.
Thanks @juncevich for working on this. Please check CI run for failures: https://github.com/juncevich/ozone/actions/runs/11940547308
| NodeSchemaManager cachedInstance = cache.get(schemaFile); | ||
| if (cachedInstance == null) { | ||
| instance = new NodeSchemaManager(); | ||
| instance.init(schemaFile); | ||
| cache.put(schemaFile, instance); | ||
| return instance; | ||
| } else { | ||
| return cachedInstance; | ||
| } |
There was a problem hiding this comment.
This is not thread-safe. Use cache.computeIfAbsent() if NodeSchemaManager.init() has no side-effects.
There was a problem hiding this comment.
Thanks, will fix.
|
Looking at stack trace in the screenshot attached to HDDS-11764, I think this is fixed by #6434. |
|
Yep. It looks like you're right. |
|
Will test and this PR close if the problem will solved. |
|
We found a performance problem with the recording file's speed. Without this fix, it degraded by approximately 14 %. We need time to find the problem. |
2 participants
HDDS-11764. Fix initiating DocumentBuilderFactory performance problem.
Cache initiating DocumentBuilderFactory and NodeSchemaManager to ip
Link to Jira task: HDDS-11764
Patch tested only by performance tests.