IMPALA-12929: Skip loading HDFS permissions in local-catalog mode

HDFS file/dir permissions are not used at all in local catalog mode - in
LocalFsTable, hasWriteAccessToBaseDir() always returns true and
getFirstLocationWithoutWriteAccess() always returns null.

However, in catalogd, we still load them (in single thread for a table!)
which could dominant the table loading time when there are lots of
partitions. Note that the table loading process in catalogd is the same
no matter what catalog mode is in used. The difference between catalog
modes is mainly in how coordinators get metadata from catalogd. Local
catalog mode is turned on by setting --catalog_topic_mode=minimal on
catalogd and --use_local_catalog=true on coordinators.

This patch skips loading HDFS permissions on catalogd when running in
local catalog mode. We can revisit it in IMPALA-7539.

Tests:
 - Ran CORE tests

Change-Id: I5baa9f6ab0d3888a78ff161ae5caa19e85bc983a
Reviewed-on: http://gerrit.cloudera.org:8080/21178
Reviewed-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
Tested-by: Impala Public Jenkins <impala-public-jenkins@cloudera.com>
This commit is contained in:
stiga-huang
2024-03-21 14:37:33 +08:00
committed by Impala Public Jenkins
parent 74b6df7997
commit 0d49c9d6cc
2 changed files with 8 additions and 0 deletions

View File

@@ -877,6 +877,10 @@ public class HdfsTable extends Table implements FeFsTable {
* {@link #getAvailableAccessLevel(String, Path, FsPermissionCache)}
*/
private static boolean assumeReadWriteAccess(FileSystem fs) {
// Avoid loading permissions in local catalog mode since they are not used in
// LocalFsTable. Remove this once we resolve IMPALA-7539.
if (BackendConfig.INSTANCE.isMinimalTopicMode()) return true;
// Avoid calling getPermissions() on file path for S3 files, as that makes a round
// trip to S3. Also, the S3A connector is currently unable to manage S3 permissions,
// so for now it is safe to assume that all files(objects) have READ_WRITE

View File

@@ -487,4 +487,8 @@ public class BackendConfig {
public String queryLogTableName() {
return backendCfg_.query_log_table_name;
}
public boolean isMinimalTopicMode() {
return backendCfg_.catalog_topic_mode.equalsIgnoreCase("minimal");
}
}