3

long time lurker, I just launched a new m1.large instance on EC2 and I need all the available instance storage on the machine(850GB) for data processing.

I understand that generally all of the memory doesn't come mounted and you have to stitch a couple drives together(generally /dev/sdb, /dev/sdc). Amazon ec2 - how to setup? However on this instance those drives aren't listed in /dev on my box...

ubuntu@ip-***:/dev$ ls
autofs           fd    hvc7   loop6               port   ram13  ram9      tty0   tty18  tty27  tty36  tty45  tty54  tty63      ttyS13  ttyS22  ttyS31   vcs    vcsa3
block            full  input  loop7               ppp    ram14  random    tty1   tty19  tty28  tty37  tty46  tty55  tty7       ttyS14  ttyS23  ttyS4    vcs1   vcsa4
btrfs-control    fuse  kmsg   loop-control        psaux  ram15  rfkill    tty10  tty2   tty29  tty38  tty47  tty56  tty8       ttyS15  ttyS24  ttyS5    vcs2   vcsa5
char             hvc0  log    mapper              ptmx   ram2   shm       tty11  tty20  tty3   tty39  tty48  tty57  tty9       ttyS16  ttyS25  ttyS6    vcs3   vcsa6
console          hvc1  loop0  mem                 pts    ram3   snapshot  tty12  tty21  tty30  tty4   tty49  tty58  ttyprintk  ttyS17  ttyS26  ttyS7    vcs4   vga_arbiter
core             hvc2  loop1  net                 ram0   ram4   snd       tty13  tty22  tty31  tty40  tty5   tty59  ttyS0      ttyS18  ttyS27  ttyS8    vcs5   xvda1
cpu              hvc3  loop2  network_latency     ram1   ram5   stderr    tty14  tty23  tty32  tty41  tty50  tty6   ttyS1      ttyS19  ttyS28  ttyS9    vcs6   xvdb
cpu_dma_latency  hvc4  loop3  network_throughput  ram10  ram6   stdin     tty15  tty24  tty33  tty42  tty51  tty60  ttyS10     ttyS2   ttyS29  uinput   vcsa   zero
disk             hvc5  loop4  null                ram11  ram7   stdout    tty16  tty25  tty34  tty43  tty52  tty61  ttyS11     ttyS20  ttyS3   urandom  vcsa1
ecryptfs         hvc6  loop5  oldmem              ram12  ram8   tty       tty17  tty26  tty35  tty44  tty53  tty62  ttyS12     ttyS21  ttyS30  usbmon0  vcsa2

As you can see there is no /dev/sdb,sdc The ebs backed drive is /dev/xvda1 and the currently mounted ephemeral/instance drive is xvdb, but there's no second ephemeral drive to mount that I can see.

df -ah
Filesystem      Size  Used Avail Use% Mounted on
/dev/xvda1      8.0G  857M  6.8G  12% /
proc               0     0     0    - /proc
sysfs              0     0     0    - /sys
none               0     0     0    - /sys/fs/fuse/connections
none               0     0     0    - /sys/kernel/debug
none               0     0     0    - /sys/kernel/security
udev            3.7G  8.0K  3.7G   1% /dev
devpts             0     0     0    - /dev/pts
tmpfs           1.5G  156K  1.5G   1% /run
none            5.0M     0  5.0M   0% /run/lock
none            3.7G     0  3.7G   0% /run/shm
/dev/xvdb       414G  199M  393G   1% /mnt

$ mount
/dev/xvda1 on / type ext4 (rw)
proc on /proc type proc (rw,noexec,nosuid,nodev)
sysfs on /sys type sysfs (rw,noexec,nosuid,nodev)
none on /sys/fs/fuse/connections type fusectl (rw)
none on /sys/kernel/debug type debugfs (rw)
none on /sys/kernel/security type securityfs (rw)
udev on /dev type devtmpfs (rw,mode=0755)
devpts on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=0620)
tmpfs on /run type tmpfs (rw,noexec,nosuid,size=10%,mode=0755)
none on /run/lock type tmpfs (rw,noexec,nosuid,nodev,size=5242880)
none on /run/shm type tmpfs (rw,nosuid,nodev)
/dev/xvdb on /mnt type ext3 (rw,_netdev)

Does anyone know why there isn't another drive I can mount and then stitch into a full 850GB disk? Or what it's labeled for that matter?

This is my first time putting a large together as well but the different drive names are throwing me for a loop and making me think I've forgotten something. Is this a change in 12.04 that I missed?
Thanks for any help!

Community
  • 1
Ropes
  • 33
  • 3
  • 2
    Most EBS-root instances attach only one (if any) ephemeral disks by default. You need to modify the `--block-device-mapping` parameter that you pass to `ec2-run-instances`. See [this answer](http://serverfault.com/a/386350/86472) for more details. – cyberx86 May 07 '12 at 20:29

1 Answers1

2

This can't be done from the GUI, but you can attach them via the command-line tools. 

ec2-run-instances ami-abcd1234 -t m1.large -b sdb=ephemeral0 -b sdc=ephemeral1 [...]

The key there is the -b command, as that tells ec2-run-instances how to set up the block-mappings. If you had several EBS volumes to attach to it, you'd do that there. For the instance-local storage, you need to explicitly declare their mappings to the instance when you create it.

Once you have two, you can do with them as you will.

sysadmin1138
  • 131,083
  • 18
  • 173
  • 296
  • Thanks for the tip! This is what I tried to run ec2-run-instances ami-a29943cb -g default -k XXXXXX -t m1.large --availability-zone us-east-1b -b xvdb=ephemeral0 -b xvdc=ephemeral1 However it attempts to start the box and then it's status goes strait to terminated. Any thoughts? – Ropes May 07 '12 at 21:50
  • @Ropes No thoughts here. Never seen that before. – sysadmin1138 May 07 '12 at 21:54
  • 1
    @Ropes: a) check the console log for errors b) I would guess, however, that the issue is that you need to pass a device (with `/dev`) to `--block-device-mapping`, e.g. `-b /dev/sdc=ephemeral1`. Also, somewhat confusingly, for Linux instances, the device you specify with `--block-device-mapping` is of the form `/dev/sdX` (not `/dev/xvdX` - which is used for Windows instances), but the device will show up on the instance as `/dev/xvdX` (with a symlink to `/dev/sdX`). The [AWS docs](http://docs.amazonwebservices.com/AWSEC2/latest/UserGuide/InstanceStorage.html) are a good reference on the topic. – cyberx86 May 08 '12 at 00:00
  • It appears that the issues were amazon/ubuntu12.04 related, I tried the same command today and it worked with a 11.10 ami. Thanks for your help! – Ropes May 08 '12 at 18:53