comnetwork commented on code in PR #4882:
URL: https://github.com/apache/hbase/pull/4882#discussion_r1023659230
##########
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HStore.java:
##########
@@ -961,10 +961,10 @@ public List<KeyValueScanner> getScanners(boolean
cacheBlocks, boolean usePread,
storeFilesToScan =
this.storeEngine.getStoreFileManager().getFilesForScan(startRow,
includeStartRow, stopRow, includeStopRow);
memStoreScanners = this.memstore.getScanners(readPt);
+ storeFilesToScan.stream().forEach(f->f.getReader().incrementRefCount());
Review Comment:
I think one problem of moving the getting scanner code into the read lock
section is that in `StoreFileScanner.getScannersForStoreFiles `method , it
would call `HStoreFile.initReader` , which would make the
`storeEngine.readLock `somewhat long time? Another option is that since
`StoreFileReader.refCount` comes from `StoreFileInfo.refCount`, we can increase
the StoreFileInfo.refCount by `HStoreFile.fileInfo`?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
