火花EC2 --ebs-VOL大小不工作 [英] spark-ec2 --ebs-vol-size not working
问题描述
在启动与火花EC2
火花集群中, - EBS-体积尺寸
标记出现没有任何效果。与 50
或 500
设置,然后ssh'ing到主节点, DF -h
显示有关空间10G的 /
。
When launching a spark cluster with spark-ec2
, the --ebs-vol-size
flag appears to have no effect. Setting it with 50
or 500
and then ssh'ing into the master node, a df -h
shows about 10G of space on /
.
如何使用火花EC2
以创建一个更大的EC2虚拟机?
How can I use spark-ec2
to create a larger EC2 virtual machine?
推荐答案
步骤有点复杂的名单已经低于提供给我工作 -
A bit elaborate list of steps have been provided below that worked for me -
- 启动一个与--ebs-VOL大小火花EC2集群
-
在./ephemeral-hdfs关机的hadoop
- Launch a spark-ec2 cluster with --ebs-vol-size
Shutdown hadoop on ./ephemeral-hdfs
./短暂-HDFS /斌/ stop-all.sh
./ephemeral-hdfs/bin/stop-all.sh
开始在./persistent-hdfs的hadoop
Start hadoop on ./persistent-hdfs
./持久HDFS /斌/ start-all.sh
./persistent-hdfs/bin/start-all.sh
您可以验证目前的规模还没有反映所请求的EBS卷大小
You can verify that the current size has not reflected the requested ebs vol size
./持久HDFS /斌/ Hadoop的dfsadmin -report
./persistent-hdfs/bin/hadoop dfsadmin -report
运行以下命令(建议把它们放到一个脚本),并运行它 -
Run the following commands ( recommend to put them into a script ) and run it -
./持久HDFS /斌/ stop-all.sh
./persistent-hdfs/bin/stop-all.sh
SED -i的#体积/持久HDFS#VOL0 /持久HDFS#G'〜/持久HDFS / conf目录/核心的site.xml
sed -i 's#vol/persistent-hdfs#vol0/persistent-hdfs#g' ~/persistent-hdfs/conf/core-site.xml
./火花EC2 / copy-dir.sh〜/持久HDFS / conf目录/核心的site.xml
./spark-ec2/copy-dir.sh ~/persistent-hdfs/conf/core-site.xml
./火花EC2 / copy-dir.sh〜/持久HDFS / conf目录/ HDFS-site.xml中
./spark-ec2/copy-dir.sh ~/persistent-hdfs/conf/hdfs-site.xml
./持久HDFS /斌/ Hadoop的NameNode的-format
./persistent-hdfs/bin/hadoop namenode -format
./持久HDFS /斌/ start-all.sh
./persistent-hdfs/bin/start-all.sh
重复步骤4,以验证大小
Repeat Step 4 to verify size
信贷 -
通过brendancol
这篇关于火花EC2 --ebs-VOL大小不工作的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!