apache/hadoop · error · NotInMountpointException

getDefaultBlockSize on path `<f>' is not within a mount poin

Error message

getDefaultBlockSize on path `<f>' is not within a mount point

What it means

ViewFileSystem.getDefaultBlockSize(Path) first resolves the path through the mount table (fsState.resolve). When resolution ends in a FileNotFoundException - the path is not covered by any configured mount link - the code translates it into NotInMountpointException("getDefaultBlockSize", f), an UnsupportedOperationException. The block size of a path is a property of the target file system, so viewfs cannot answer for unmounted paths.

Source

Thrown at hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/viewfs/ViewFileSystem.java:973

  @Override
  public short getDefaultReplication() {
    throw new NotInMountpointException("getDefaultReplication");
  }

  @Override
  public FsServerDefaults getServerDefaults() throws IOException {
    throw new NotInMountpointException("getServerDefaults");
  }

  @Override
  public long getDefaultBlockSize(Path f) {
    try {
      InodeTree.ResolveResult<FileSystem> res =
        fsState.resolve(getUriPath(f), true);
      return res.targetFileSystem.getDefaultBlockSize(res.remainingPath);
    } catch (FileNotFoundException e) {
      throw new NotInMountpointException(f, "getDefaultBlockSize");
    } catch (IOException e) {
      throw new RuntimeException("Not able to initialize fs in "
          + " getDefaultBlockSize for path " + f + " with exception", e);
    }
  }

  @Override
  public short getDefaultReplication(Path f) {
    try {
      InodeTree.ResolveResult<FileSystem> res =
        fsState.resolve(getUriPath(f), true);
      return res.targetFileSystem.getDefaultReplication(res.remainingPath);
    } catch (FileNotFoundException e) {
      throw new NotInMountpointException(f, "getDefaultReplication");
    } catch (IOException e) {
      throw new RuntimeException("Not able to initialize fs in "
          + " getDefaultReplication for path " + f + " with exception", e);
    }

View on GitHub (pinned to 2add963021)

Solutions

1. Add the missing mount entry, e.g. fs.viewfs.mounttable.default.link.tmp=/tmp on the target cluster in core-site.xml
2. Call getDefaultBlockSize on a path that is already known to be mounted (verify with `hadoop fs -ls viewfs://cluster/`)
3. Add fs.viewfs.mounttable.default.linkMergeSlash=hdfs://nameservice1/ so every path resolves to a real file system
4. Catch NotInMountpointException in generic tooling and fall back to the target cluster's dfs.blocksize

Example fix

<!-- before: /tmp not mounted, getDefaultBlockSize(new Path("/tmp")) throws -->
<property>
  <name>fs.viewfs.mounttable.default.link.data</name>
  <value>hdfs://nameservice1/data</value>
</property>

<!-- after: add the tmp mount -->
<property>
  <name>fs.viewfs.mounttable.default.link.tmp</name>
  <value>hdfs://nameservice1/tmp</value>
</property>

// Verify the path prefix is covered by a configured mount before asking for block size
static boolean isMounted(Configuration conf, Path p) {
  String path = Path.getPathWithoutSchemeAndAuthority(p).toString();
  if (conf.get("fs.viewfs.mounttable.default.linkMergeSlash") != null) return true;
  for (Map.Entry<String,String> e : conf) {
    String k = e.getKey();
    if (k.startsWith("fs.viewfs.mounttable.default.link.")
        && ("/" + path).startsWith(k.substring("fs.viewfs.mounttable.default.link.".length()))) {
      return true;
    }
  }
  return false;
}

try {
  long bs = fs.getDefaultBlockSize(p);
} catch (NotInMountpointException e) {
  // p not covered by the mount table: mount it or use cluster default
  long bs = conf.getLongBytes("dfs.blocksize", 128L * 1024 * 1024);
}

When it happens

Trigger: Calling getDefaultBlockSize(new Path("/tmp")) when no fs.viewfs.mounttable.default.link.tmp entry exists; calling it on "/" or on a mount-table-only internal directory (e.g. /data when only /data/hdfs is mounted); typos in the path that fall outside every mount prefix.

Common situations: Cluster migration to viewfs where the old absolute paths (/tmp, /user, /data) were not all added to the mount table; jobs assuming /tmp exists; mount table loaded from a config file that is not on the client classpath so the table is effectively empty.

Related errors

• ViewFs: Non absolute mount name in config:{src}
• Path {nextInode.fullPath} already exists as link
• Path {strB} already exists as dir; cannot create link here
• Unexpected mount table link entry '{key}'. Use linkMergeSlas
• ViewFs: Cannot initialize: Invalid entry in Mount table in c

AI-assisted analysis of apache/hadoop@2add963021 (2026-08-22). Data as JSON: /api/errors/101107b081880eeb. Report an issue: GitHub.