All geek questions in one place

Hadoop on Amazon EC2: Job Tracker Doesn't Start Properly

We run Hadoop on an Amazon EC2 cluster. We start the wizard, the slaves and attach the ebs volumes and finally wait for the jobtracker hadoop, tasktracker, etc. to start, and we have a wait time of 3600 seconds. We notice 50% of the time when the tracker cannot start working before the timeout. The reason is that hdfs is not initializing properly and is still in safemode mode and the job tracker cannot start. I noticed several connectivity issues between nodes on EC2 when I tried to manually check the slaves.

Has anyone encountered a similar problem and knew how to solve this?

+3
source share
1 answer

I'm not sure if this issue is related to Amazon EC2. I also had this problem, although there was a pseudo-distributed installation on my machine.

In these cases, I can disable safe mode manually and safely.
Try this command:bin/hadoop dfsadmin -safemode leave

I think you cannot be wrong here. This seems to be a chaotic feature of howoop. I used 0.18.3, which version are you using?

0
source

Source: https://habr.com/ru/post/1744097/

More articles:


All Articles