1) on the virtual machine installed Ubuntu Desktop version as Datanode, because the physical machine memory limit only divided 1G of memory to the virtual machine, using bin/start-all.sh started Hadoop, slave resource usage as shown:
A total of 986M of memory was used for 574M, which has not yet started to run the Mr Program.
2) Workaround: Set to enable Ubuntu to boot directly into the command line mode, rather than desktop mode. Use the following command to set up
sudo gedit/etc/default/grub
Find this line
grub_cmdline_linux_default="quiet Splash"
Change into
grub_cmdline_linux_default="quiet Splash text"
Re-enter the command:
sudo update-grub
It automatically enters the command line (tty1) after booting.
In addition, sometimes you may need to enter the desktop mode, command-line mode needs to return to the graphics mode, input command: STARTX
If you want to switch back again, just restart it, because this can happen with CTRL+ALT+F1:
3) Finally, using command-line mode, when running Hadoop, memory usage is as follows:
When command-line mode is turned on, memory usage is reduced to 235M, saving 300+m.
Resolves low memory---as slave virtual machine under Ubuntu