ChenSammi commented on code in PR #3292:
URL: https://github.com/apache/ozone/pull/3292#discussion_r849190788
##########
hadoop-hdds/container-service/src/main/java/org/apache/hadoop/ozone/container/common/utils/HddsVolumeUtil.java:
##########
@@ -234,4 +257,154 @@ public static boolean checkVolume(HddsVolume hddsVolume,
String scmId,
return true;
}
}
+
+ /**
+ * Randomly pick a DbVolume for HddsVolume and init db instance.
+ * Use the HddsVolume directly if no DbVolume found.
+ * @param hddsVolume
+ * @param dbVolumeSet
+ * @param clusterID
+ * @param conf
+ */
+ public static void formatDbStoreForHddsVolume(HddsVolume hddsVolume,
+ MutableVolumeSet dbVolumeSet, String clusterID,
+ ConfigurationSource conf) throws IOException {
+ DbVolume chosenDbVolume = null;
+ File dbParentDir;
+
+ if (dbVolumeSet == null || dbVolumeSet.getVolumesList().isEmpty()) {
+ // No extra db volumes specified, just create db under the HddsVolume.
+ dbParentDir = new File(hddsVolume.getStorageDir(), clusterID);
+ } else {
+ // Randomly choose a DbVolume for simplicity.
+ List<DbVolume> dbVolumeList = StorageVolumeUtil.getDbVolumesList(
+ dbVolumeSet.getVolumesList());
+ chosenDbVolume = dbVolumeList.get(
+ ThreadLocalRandom.current().nextInt(dbVolumeList.size()));
+ dbParentDir = new File(chosenDbVolume.getStorageDir(), clusterID);
+ }
+
+ // Init subdir with the storageID of HddsVolume.
+ File storageIdDir = new File(dbParentDir, hddsVolume.getStorageID());
+ if (!storageIdDir.mkdirs() && !storageIdDir.exists()) {
+ throw new IOException("Can't make subdir under "
+ + dbParentDir.getAbsolutePath() + " for volume "
+ + hddsVolume.getStorageID());
+ }
+
+ // Init the db instance for HddsVolume under the subdir above.
+ String containerDBPath = new File(storageIdDir, CONTAINER_DB_NAME)
+ .getAbsolutePath();
+ try {
+ initPerDiskDBStore(containerDBPath, conf);
+ } catch (IOException e) {
+ throw new IOException("Can't init db instance under path "
+ + containerDBPath + " for volume " + hddsVolume.getStorageID());
+ }
+
+ // Set the dbVolume and dbParentDir of the HddsVolume for db path lookup.
+ hddsVolume.setDbVolume(chosenDbVolume);
+ hddsVolume.setDbParentDir(storageIdDir);
+ }
+
+ /**
+ * Load already formatted db instances for all HddsVolumes.
+ * @param hddsVolumeSet
+ * @param dbVolumeSet
+ * @param conf
+ * @param logger
+ */
+ public static void loadAllHddsVolumeDbStore(MutableVolumeSet hddsVolumeSet,
+ MutableVolumeSet dbVolumeSet, ConfigurationSource conf, Logger logger) {
+ // Scan subdirs under the db volumes and build a one-to-one map
+ // between each HddsVolume -> DbVolume.
+ mapDbVolumesToDataVolumesIfNeeded(hddsVolumeSet, dbVolumeSet);
+
+ List<HddsVolume> hddsVolumeList = StorageVolumeUtil.getHddsVolumesList(
+ hddsVolumeSet.getVolumesList());
+
+ for (HddsVolume volume : hddsVolumeList) {
+ String clusterID = volume.getClusterID();
+
+ // DN startup for the first time, not registered yet,
+ // so the DbVolume is not formatted.
+ if (clusterID == null) {
+ continue;
+ }
+
+ File clusterIdDir = new File(volume.getDbVolume() == null ?
+ volume.getStorageDir() : volume.getDbVolume().getStorageDir(),
+ clusterID);
+
+ if (!clusterIdDir.exists()) {
+ if (logger != null) {
+ logger.error("HddsVolume {} db instance not formatted, " +
+ "clusterID {} directory not found",
+ volume.getStorageDir().getAbsolutePath(),
+ clusterIdDir.getAbsolutePath());
+ }
+ continue;
Review Comment:
It depends on how many disk failure is tolerated. For example, in HDFS,
default disk failure tolerated is 0, which means as long as one disk fails, DN
cannot start up. Admin may change the configuration based on his/her
estimation of disk failure rate. That's acceptable and the risk is on Admin.
So "at least one disk volume" is not acceptable, the risk is too high.
Think about this case, all HDDS volumes function well while one of the two
configured dbVolumes is down at DN startup. How will the impacted HDDS volume
behave? Create a new RocksDB instance on the remaining dbVolume and then go on
to provide service? I think we should persist this HDDS volume to RocksDB
instance relation in HDDS version file.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
