guluo2016 commented on PR #5767:
URL: https://github.com/apache/hbase/pull/5767#issuecomment-2016469719
> Mind explaining more on why locating is so slow when the replica id does
not exist?
Busying with other things, Sorry for the late reply.
In a short, hbase would retry 8 times (by default) if the location of
specified region is not obtained, and this stage take a long time.
A detailed analysis for this situation against master branch
We will get RegionOfflineException if region replica id does not exist. Code
is in here
```
// AsyncRegionLocator.getRegionLocation
// loc == null because region replica id does not exist
HRegionLocation loc = locs.getRegionLocation(replicaId);
if (loc == null) {
future.completeExceptionally(
new RegionOfflineException("No location for " + tableName + ", row='"
+ Bytes.toStringBinary(row) + "', locateType=" + type + ",
replicaId=" + replicaId));
}
```
HBase would retry it over and over util the maxAttempts is reached when
getting RegionOfflineException because this exception is not
DoNotRetryIOException, Code is in here.
```
// AsyncRpcRetryingCaller.onError
if (error instanceof DoNotRetryIOException && !(error instanceof
ScannerResetException)) {
future.completeExceptionally(error);
return;
}
...
tryScheduleRetry(error);
```
We are sure that the region does not exist in this situation, so no need to
retry maybe is better.
So we can throw DoNotRetryIOException to avoid retrying in this situation
--
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: issues-unsubscr...@hbase.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
