Local Storage Datastore Configuration

Community Support
1

Hi all,

I was wondering if someone could clarify the configuration process of setting up a local HDD datastore?

We are exploring a tiered storage architecture with Ceph. However, in some cases, the workload is better suited to a local SSD/HDD on a particular host.

I’ve setup a ‘System’ target for ‘Images’ on one of my hosts. I setup the base path to be /tmp and it appears to be getting capacity information correctly.

Here are the attributes of the datastore I setup for local use:

BASE_PATH /tmp/
BRIDGE_LIST 10.175.71.160
CLONE_TARGET SYSTEM
DISK_TYPE FILE
DS_MAD fs
LN_TARGET SYSTEM
TM_MAD ssh
TYPE IMAGE_DS

Logs from the VM I tried to launch:

Tue Oct 13 17:15:10 2015 [Z0][DiM][I]: New VM state is ACTIVE.
Tue Oct 13 17:15:13 2015 [Z0][LCM][I]: New VM state is PROLOG.
Tue Oct 13 17:15:14 2015 [Z0][TM][I]: Command execution fail: /var/lib/one/remotes/tm/ssh/clone opennebula-01:/tmp/103/923a58161f57ef623849e7f74cf4559c 10.175.71.160:/var/lib/one//datastores/0/2244/disk.0 2244 103
Tue Oct 13 17:15:14 2015 [Z0][TM][I]: clone: Cloning opennebula-01:/tmp/103/923a58161f57ef623849e7f74cf4559c in /var/lib/one/datastores/0/2244/disk.0
Tue Oct 13 17:15:14 2015 [Z0][TM][E]: clone: Command “scp -r opennebula-01:/tmp/103/923a58161f57ef623849e7f74cf4559c 10.175.71.160:/var/lib/one//datastores/0/2244/disk.0” failed: Warning: Permanently added ‘opennebula-01,10.175.71.120’ (ECDSA) to the list of known hosts.
Tue Oct 13 17:15:14 2015 [Z0][TM][I]: Warning: Permanently added ‘10.175.71.160’ (ECDSA) to the list of known hosts.
Tue Oct 13 17:15:14 2015 [Z0][TM][I]: /tmp/103/923a58161f57ef623849e7f74cf4559c: No such file or directory
Tue Oct 13 17:15:14 2015 [Z0][TM][E]: Error copying opennebula-01:/tmp/103/923a58161f57ef623849e7f74cf4559c to 10.175.71.160:/var/lib/one//datastores/0/2244/disk.0
Tue Oct 13 17:15:14 2015 [Z0][TM][I]: ExitCode: 1
Tue Oct 13 17:15:14 2015 [Z0][TM][E]: Error executing image transfer script: Error copying opennebula-01:/tmp/103/923a58161f57ef623849e7f74cf4559c to 10.175.71.160:/var/lib/one//datastores/0/2244/disk.0
Tue Oct 13 17:15:15 2015 [Z0][DiM][I]: New VM state is FAILED

Thanks for any clarification that can be offered!

2

Hi Jeff

It seems it’s complaining about a missing file, could you check that both the original file (opennebula-01:/tmp/103/923a58161f57ef623849e7f74cf4559c) and target dir 10.175.71.160:/var/lib/one//datastores/0/2244 exists?

Cheers

3

Hi Ruben,

The directory exists in /var/lib/one and /tmp.

Thanks!

4

I had the same problem. is there some solution?

5

Hi,

I would like just to clarify:

opennebula-01 is the host where the front end (sunstone) is working?

what do you get if you run as oneadmin (on the frontend)

su - oneadmin
ssh opennebula-01 ls -la /tmp/103/923a58161f57ef623849e7f74cf4559c

Cheers,
Anton Todorov

6

I created one datastore, type IMAGE, TM_MAD ssh. I don’t created other datastore type system. My goal is to use node’s space and not frontend. Soul, I try with TM_MAD ssh and BRIDGE_LIST my node(host1) IP. Everything OK when I downloaded images in Marketplace and created templates, but when I try create virtual machines using that templates occurs errors below:

Tue Mar 15 18:59:48 2016 [Z0][VM][I]: New state is ACTIVE
Tue Mar 15 18:59:48 2016 [Z0][VM][I]: New LCM state is PROLOG
Tue Mar 15 18:59:49 2016 [Z0][TM][I]: Command execution fail: /var/lib/one/remotes/tm/ssh/clone frontend:/storage/105/49dc9f4ade78bbcc6d858d71214f7300 host1:/var/lib/one//datastores/0/29/disk.0 29 105
Tue Mar 15 18:59:49 2016 [Z0][TM][I]: clone: Cloning frontend:/storage/105/49dc9f4ade78bbcc6d858d71214f7300 in /var/lib/one/datastores/0/29/disk.0
Tue Mar 15 18:59:49 2016 [Z0][TM][E]: clone: Command “scp -r frontend:/storage/105/49dc9f4ade78bbcc6d858d71214f7300 host1:/var/lib/one//datastores/0/29/disk.0” failed: Warning: Permanently added ‘frontend’ (ECDSA) to the list of known hosts.
Tue Mar 15 18:59:49 2016 [Z0][TM][I]: Warning: Permanently added ‘host1,192.5.0.40’ (ECDSA) to the list of known hosts.
Tue Mar 15 18:59:49 2016 [Z0][TM][I]: /storage/105/49dc9f4ade78bbcc6d858d71214f7300: No such file or directory
Tue Mar 15 18:59:49 2016 [Z0][TM][E]: Error copying frontend:/storage/105/49dc9f4ade78bbcc6d858d71214f7300 to host1:/var/lib/one//datastores/0/29/disk.0
Tue Mar 15 18:59:49 2016 [Z0][TM][I]: ExitCode: 1
Tue Mar 15 18:59:49 2016 [Z0][TM][E]: Error executing image transfer script: Error copying frontend:/storage/105/49dc9f4ade78bbcc6d858d71214f7300 to host1:/var/lib/one//datastores/0/29/disk.0
Tue Mar 15 18:59:49 2016 [Z0][VM][I]: New LCM state is PROLOG_FAILURE

Obs:

  • My infrastructure is: one frontend and one node(host1).
  • /var/lib/one is default datastore by examples (NFS).
  • /storage is share in Host1.
  • frontend log in node1 without password ssh.

I need to create virtual machines using /storage (node1) and not /var/lib/one (frontend).

Someone can I help me?
Thanks.