namenode goes in safemode every time i re-start machine

Asked Feb 20 '18 at 19:00

Active Feb 20 '18 at 19:09

Viewed 408 times

I have Hadoop 1.x on ubuntu

Everything works fine: I'm able to upload input file in HDFS. But when I reboot my laptop, all HDFS blocks are corrupted and NameNode starts in SafeMode

So I have to

1) Leave the SafeMode

2) Delete all corrupted blocks with

hdfs fsck -delete

3)Re-upload input file

Then it works fine until next reboot.

Can someone please get me some solution for this. Thanks

edited Feb 20 '18 at 19:09

asked Feb 20 '18 at 19:00

Ankit dhoot

I think your first problem is using Hadoop 1.x... Never reboot your machine with cleaning stopping HDFS, otherwise, yes, it'll corrupt the open files – OneCricketeer Feb 21 '18 at 01:42
but what is the solution for this problem ? how can i over come with this namenode safemode problem and data corrupt?? – Ankit dhoot Feb 21 '18 at 05:33
`$HADOOP_HOME/bin/stop-dfs.sh`, wait for it, and then turn off your computer? – OneCricketeer Feb 21 '18 at 05:34
still data is lost when i run stop-dfs.sh then restart laptop.. is there any need to add properties in hdfs file? @cricket_007 – Ankit dhoot Feb 21 '18 at 05:42
No data should be lost when you gracefully stop HDFS using that command. You might have to stop the datanode process as well. – OneCricketeer Feb 21 '18 at 05:47
still when i re start machine then namenode goes permanently in safe mode and data is corrupted. – Ankit dhoot Feb 21 '18 at 13:17

namenode goes in safemode every time i re-start machine

0 Answers0