Apache
/
hadoop
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

HDFS-15644. Failed volumes can cause DNs to stop block reporting. Contributed by Ahmed Hussein.

This commit is contained in:
Wei-Chiu Chuang 2020-10-28 09:22:47 -07:00
parent 90ebbaa393
commit e7e165a927
1 changed files with 22 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/datanode/fsdataset/impl/FsDatasetImpl.java
View File

@ -1935,29 +1935,33 @@ class FsDatasetImpl implements FsDatasetSpi<FsVolumeImpl> {
Set<String> missingVolumesReported = new HashSet<>();
for (ReplicaInfo b : volumeMap.replicas(bpid)) {
String volStorageID = b.getVolume().getStorageID();
if (!builders.containsKey(volStorageID)) {
switch(b.getState()) {
case FINALIZED:
case RBW:
case RWR:
break;
case RUR:
// use the original replica.
ReplicaUnderRecovery rur = (ReplicaUnderRecovery) b;
b = rur.getOriginalReplica();
break;
case TEMPORARY:
continue;
default:
assert false : "Illegal ReplicaInfo state.";
continue;
}
BlockListAsLongs.Builder storageBuilder = builders.get(volStorageID);
// a storage in the process of failing will not be in the volumes list
// but will be in the replica map.
if (storageBuilder != null) {
storageBuilder.add(b);
} else {
if (!missingVolumesReported.contains(volStorageID)) {
LOG.warn("Storage volume: " + volStorageID + " missing for the"
+ " replica block: " + b + ". Probably being removed!");
missingVolumesReported.add(volStorageID);
}
continue;
}
switch(b.getState()) {
case FINALIZED:
case RBW:
case RWR:
builders.get(b.getVolume().getStorageID()).add(b);
break;
case RUR:
ReplicaUnderRecovery rur = (ReplicaUnderRecovery)b;
builders.get(rur.getVolume().getStorageID())
.add(rur.getOriginalReplica());
break;
case TEMPORARY:
break;
default:
assert false : "Illegal ReplicaInfo state.";
}
}
}