Bug 1358116 - Topology load fails if the disk sizes are of different sizes
Summary: Topology load fails if the disk sizes are of different sizes
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: heketi
Version: unspecified
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
: ---
Assignee: Luis Pabón
QA Contact: Jonathan Holloway
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2016-07-20 06:14 UTC by Bhaskarakiran
Modified: 2016-11-23 23:13 UTC (History)
5 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2016-10-05 21:52:10 UTC
Embargoed:


Attachments (Terms of Use)
Heketi log (200.92 KB, text/plain)
2016-07-21 07:22 UTC, Bhaskarakiran
no flags Details

Description Bhaskarakiran 2016-07-20 06:14:18 UTC
Description of problem:
======================
Tried to create a 3 node topology with one node having 100G and another 2 nodes having 500G. topology creation fails after one node. If the 1st node is added with 500G then it works. Ideally it should not be checking the disk sizes to be same on all the nodes. This might be a problem with dynamic provisioning.

Version-Release number of selected component (if applicable):
=============================================================
Heketi 2.0.5

How reproducible:
100%

Steps to Reproduce:
1. Create a topology with 3 nodes having different disk sizes.
2.
3.

Actual results:
===============
Topology creation fails.

Expected results:
=================
Should pass

Additional info:

Comment 3 Luis Pabón 2016-07-20 20:11:59 UTC
Heketi does not check disk size.  I think something else is going on.

Please attach the heketi logs

Comment 4 Luis Pabón 2016-07-20 20:13:17 UTC
Please attach Heketi logs.  If you are using OpenShift, you can type:

oc logs pod <heketi-pod> > heketi.log

Comment 5 Luis Pabón 2016-07-21 03:45:24 UTC
Also, you may find it beneficial to try out new v2.0.6

Comment 6 Bhaskarakiran 2016-07-21 07:22:46 UTC
Created attachment 1182337 [details]
Heketi log

Attaching the heketi log.

Comment 7 Luis Pabón 2016-09-01 15:56:26 UTC
Here is the issue:

[kubeexec] ERROR 2016/07/19 11:17:10 /src/github.com/heketi/heketi/executors/kubeexec/kubeexec.go:280: Failed to run command [pvcreate --metadatasize=128M --dataalignment=256K /dev/vdb] on glusterfs-dc-dhcp41-164.lab.eng.blr.redhat.com-1-rt3vf: Err[error executing remote command: Error executing command in container: Error executing in Docker Container: 5]: Stdout []: Stderr [  Can't initialize physical volume "/dev/vdb" of volume group "vg_8d940314f90e120708cc85a2549fb778" without -ff
]

Volumes must be in raw (empty) format.  No filesystem or any other volume manager must have metadata on the disk.

Please rezero the disk:

dd if=/dev/zero of=/dev/vdb count=100 bs=1024k

Repeat for all disks to make sure they are clean.


Note You need to log in before you can comment on or make changes to this bug.