keith-turner commented on code in PR #5497:
URL: https://github.com/apache/accumulo/pull/5497#discussion_r2056364064
##########
server/tserver/src/main/java/org/apache/accumulo/tserver/log/LogSorter.java:
##########
@@ -274,12 +274,7 @@ void writeBuffer(String destPath,
List<Pair<LogFileKey,LogFileValue>> buffer, in
var logFileKey = pair.getFirst();
var logFileValue = pair.getSecond();
Key k = logFileKey.toKey();
- var list = keyListMap.putIfAbsent(k, logFileValue.mutations);
- if (list != null) {
- var muts = new ArrayList<>(list);
- muts.addAll(logFileValue.mutations);
- keyListMap.put(logFileKey.toKey(), muts);
- }
+ keyListMap.computeIfAbsent(k, (key) -> new
ArrayList<>()).addAll(logFileValue.mutations);
Review Comment:
No we do not need that optimization. The key in this case is not related to
the data, but is related to the tablet and minor compaction count. So for each
tablet in the walog it will have the same key until a minor compactoin happens
at which point the key changes. Since the key is per tablet in walog, does
not seem that further optimization to avoid the initial copy is worthwhile.
Looking into this I realized the problem is more general than what was
described in the issue. Can run into this problem even if writing different
keys to the same tablet. The problem is more of a function of how many entries
an individual tablet has in the walog, does not really matter what the data in
the tablet is.
--
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]
