HDFS fsck command shows health as corrupt for '/'

Question

I have open source hadoop version 2.7.3 cluster (2 Masters + 3 Slaves) installed on AWS EC2 instances. I am using the cluster to integrate it with Kafka Connect.

The setup of cluster was done last month and setup of kafka connect was completed last fortnight. Since then, we were able to operate the kafka topic records on our HDFS and do various operations.

Since last afternoon, I started getting below error. When I copy a new file to the cluster from local, it comes and gets opened but after some time, again starts showing similar IOException:

17/02/14 07:57:55 INFO hdfs.DFSClient: No node available for BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 file=/test/inputdata/derby.log
17/02/14 07:57:55 INFO hdfs.DFSClient: Could not obtain BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 from any node: java.io.IOException: No live nodes contain block BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 after checking nodes = [], ignoredNodes = null No live nodes contain current block Block locations: Dead nodes: . Will get new block locations from namenode and retry...
17/02/14 07:57:55 WARN hdfs.DFSClient: DFS chooseDataNode: got # 1 IOException, will wait for 499.3472970548959 msec.
17/02/14 07:57:55 INFO hdfs.DFSClient: No node available for BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 file=/test/inputdata/derby.log
17/02/14 07:57:55 INFO hdfs.DFSClient: Could not obtain BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 from any node: java.io.IOException: No live nodes contain block BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 after checking nodes = [], ignoredNodes = null No live nodes contain current block Block locations: Dead nodes: . Will get new block locations from namenode and retry...
17/02/14 07:57:55 WARN hdfs.DFSClient: DFS chooseDataNode: got # 2 IOException, will wait for 4988.873277172643 msec.
17/02/14 07:58:00 INFO hdfs.DFSClient: No node available for BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 file=/test/inputdata/derby.log
17/02/14 07:58:00 INFO hdfs.DFSClient: Could not obtain BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 from any node: java.io.IOException: No live nodes contain block BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 after checking nodes = [], ignoredNodes = null No live nodes contain current block Block locations: Dead nodes: . Will get new block locations from namenode and retry...
17/02/14 07:58:00 WARN hdfs.DFSClient: DFS chooseDataNode: got # 3 IOException, will wait for 8598.311122824263 msec.
17/02/14 07:58:09 WARN hdfs.DFSClient: Could not obtain block: BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 file=/test/inputdata/derby.log No live nodes contain current block Block locations: Dead nodes: . Throwing a BlockMissingException
17/02/14 07:58:09 WARN hdfs.DFSClient: Could not obtain block: BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 file=/test/inputdata/derby.log No live nodes contain current block Block locations: Dead nodes: . Throwing a BlockMissingException
17/02/14 07:58:09 WARN hdfs.DFSClient: DFS Read
org.apache.hadoop.hdfs.BlockMissingException: Could not obtain block: BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 file=/test/inputdata/derby.log
        at org.apache.hadoop.hdfs.DFSInputStream.chooseDataNode(DFSInputStream.java:983)
        at org.apache.hadoop.hdfs.DFSInputStream.blockSeekTo(DFSInputStream.java:642)
        at org.apache.hadoop.hdfs.DFSInputStream.readWithStrategy(DFSInputStream.java:882)
        at org.apache.hadoop.hdfs.DFSInputStream.read(DFSInputStream.java:934)
        at java.io.DataInputStream.read(DataInputStream.java:100)
        at org.apache.hadoop.io.IOUtils.copyBytes(IOUtils.java:85)
        at org.apache.hadoop.io.IOUtils.copyBytes(IOUtils.java:59)
        at org.apache.hadoop.io.IOUtils.copyBytes(IOUtils.java:119)
        at org.apache.hadoop.fs.shell.Display$Cat.printToStdout(Display.java:107)
        at org.apache.hadoop.fs.shell.Display$Cat.processPath(Display.java:102)
        at org.apache.hadoop.fs.shell.Command.processPaths(Command.java:317)
        at org.apache.hadoop.fs.shell.Command.processPathArgument(Command.java:289)
        at org.apache.hadoop.fs.shell.Command.processArgument(Command.java:271)
        at org.apache.hadoop.fs.shell.Command.processArguments(Command.java:255)
        at org.apache.hadoop.fs.shell.Command.processRawArguments(Command.java:201)
        at org.apache.hadoop.fs.shell.Command.run(Command.java:165)
        at org.apache.hadoop.fs.FsShell.run(FsShell.java:287)
        at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:70)
        at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:84)
        at org.apache.hadoop.fs.FsShell.main(FsShell.java:340)
cat: Could not obtain block: BP-1831277630-10.16.37.124-1484306078618:blk_1073793876_55013 file=/test/inputdata/derby.log

When I do : hdfs fsck / , I get:

Total size:    667782677 B
 Total dirs:    406
 Total files:   44485
 Total symlinks:                0
 Total blocks (validated):      43767 (avg. block size 15257 B)
  ********************************
  UNDER MIN REPL'D BLOCKS:      43766 (99.99772 %)
  dfs.namenode.replication.min: 1
  CORRUPT FILES:        43766
  MISSING BLOCKS:       43766
  MISSING SIZE:         667781648 B
  CORRUPT BLOCKS:       43766
  ********************************
 Minimally replicated blocks:   1 (0.0022848265 %)
 Over-replicated blocks:        0 (0.0 %)
 Under-replicated blocks:       0 (0.0 %)
 Mis-replicated blocks:         0 (0.0 %)
 Default replication factor:    3
 Average block replication:     6.8544796E-5
 Corrupt blocks:                43766
 Missing replicas:              0 (0.0 %)
 Number of data-nodes:          3
 Number of racks:               1
FSCK ended at Tue Feb 14 07:59:10 UTC 2017 in 932 milliseconds


The filesystem under path '/' is CORRUPT

That means, all my files got corrupted somehow.

I want to recover my HDFS and fix the corrupt health status. Also, I would like to understand, how such an issue occurred suddenly and how to prevent it in future?

Answer 1

The entire filesystem (43766 blocks) being marked as corrupt can be either due to removal of the dfs.datanode.data.dir folder(s) entirely or changing its value in hdfs-site.xml. Whenever doing so, make sure the Namenode is also formatted and restarted.

If not, the Namenode still holds the block information and expects them to be available among the Datanode(s). The scenario posted in the question is a similar one.

It is not possible to recover the blocks if the data directories are deleted. If only the value of dfs.datanode.data.dir property is changed in hdfs-site.xml and the Namenode is not yet formatted, reverting the value in hdfs-site.xml will help.