FAQ
Ladies and gents,

all my data nodes have concerning health, because they have swapped pages
in the last 15 minutes.
Thats the default setting for theis warning.

I have 2 cores per node and 8 GB of RAM (my piloting cluster). Do you have
any advice on how I should deal with it?

Should I upgrade RAM? How much performnace am I loosing here due to
swapping?

Any experience and advice is appreciated.

Stefan

Search Discussions

  • Philip Zeyliger at Jun 13, 2013 at 9:36 pm
    Hi Stefan,

    You can use the charts feature to figure out just how much swapping was
    actually going on. Use the timeseries query 'select dt0(swap_out) WHERE
    category=host' . If it's not too many pages, it might not be a big deal.

    We tend to recommend turning off swapping entirely. Here's how you do that:

    # minimize swapiness
    sudo sysctl vm.swappiness=0
    sudo bash -c 'echo "vm.swappiness=0" >> /etc/sysctl.conf'
    # disable swap space until next reboot:
    sudo /sbin/swapoff -a

    Cheers,

    -- Philip

    On Thu, Jun 13, 2013 at 10:06 AM, Stefan wrote:

    Ladies and gents,

    all my data nodes have concerning health, because they have swapped pages
    in the last 15 minutes.
    Thats the default setting for theis warning.

    I have 2 cores per node and 8 GB of RAM (my piloting cluster). Do you have
    any advice on how I should deal with it?

    Should I upgrade RAM? How much performnace am I loosing here due to
    swapping?

    Any experience and advice is appreciated.

    Stefan

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupscm-users @
categorieshadoop
postedJun 13, '13 at 5:06p
activeJun 13, '13 at 9:36p
posts2
users2
websitecloudera.com
irc#hadoop

2 users in discussion

Philip Zeyliger: 1 post Stefan: 1 post

People

Translate

site design / logo © 2022 Grokbase