Re: Increase /tmp space

From: Ryan G <no email>
Date: Fri, 24 Jun 2016 12:23:00 -0400

Here's what I use as a user data script. It will take all the ephemeral
storage drives and create 1 large logical volume.


#!/bin/bash

VOLUMES=""
for device in `curl -s
169.254.169.254/latest/meta-data/block-device-mapping/`
do
        if [[ $device == "ephemeral"* ]]
        then
                block=`curl -s
169.254.169.254/latest/meta-data/block-device-mapping/$device | awk -F/
'{print $NF}'`;
                if [[ -e /dev/$block ]]
                then
#if [ ! -e /mnt/$device ]
#then
               # mkfs.ext3 /dev/$block
               # mkdir /mnt/$device
               # mount /dev/$block /mnt/$device
# chmod 1777 /mnt/$device
#fi
pvcreate /dev/$block
VOLUMES="${VOLUMES} /dev/$block"
                fi
        fi
done

vgcreate vg_ephemeral $VOLUMES
SIZE=`vgdisplay vg_ephemeral | grep "Total PE" | awk '{print $3}'`
lvcreate -l $SIZE vg_ephemeral -n ephemerallv
mkfs.ext3 /dev/mapper/vg_ephemeral-ephemerallv
mkdir /scratch
mount /dev/mapper/vg_ephemeral-ephemerallv /scratch
chmod 1777 /scratch


On Fri, Jun 17, 2016 at 1:53 PM, Alex Lee <simplitia_at_gmail.com> wrote:

> Dear all,
> I notice that the AMI image I'm using is only 8G, however I'm wondering if
> there is an option to add instance storage during the launch say to 100G?
> I do have an ebs volume that I can try to remount /tmp too but I'm sure if
> that will be properly nfs shared across the nodes, also I get a feeling
> that it will much slower.
>
> here are what my devices look like at the moment.
>
> NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
> xvda 202:0 0 8G 0 disk
> └─xvda1 202:1 0 8G 0 part /
> xvdz 202:6400 0 1.5T 0 disk /data
> xvdaa 202:6656 0 300G 0 disk
>
>
> by the way what is xvdaa? Can I use it as a /tmp space?
>
> thanks in advance!
> Alex
>
> _______________________________________________
> StarCluster mailing list
> StarCluster_at_mit.edu
> http://mailman.mit.edu/mailman/listinfo/starcluster
>
>
Received on Fri Jun 24 2016 - 12:23:02 EDT