Bug 1433614 - cns-deploy fails in case there are "dos" partition signature on devices dedicated for bricks in topology file
Summary: cns-deploy fails in case there are "dos" partition signature on devices dedic...
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: heketi
Version: rhgs-3.0
Hardware: x86_64
OS: Linux
unspecified
high
Target Milestone: ---
: CNS 3.10
Assignee: Niels de Vos
QA Contact: Ashmitha Ambastha
URL:
Whiteboard: aos-scalability-35
Depends On:
Blocks: 1568861
TreeView+ depends on / blocked
 
Reported: 2017-03-18 19:52 UTC by Elvir Kuric
Modified: 2018-09-12 09:23 UTC (History)
12 users (show)

Fixed In Version: heketi-6.0.0-14.el7rhgs
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2018-09-12 09:22:12 UTC
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Github heketi heketi pull 1185 0 None None None 2018-05-16 09:47:33 UTC
Red Hat Product Errata RHEA-2018:2686 0 None None None 2018-09-12 09:23:25 UTC

Description Elvir Kuric 2017-03-18 19:52:06 UTC
Description of problem:

if in topology file device specified for brick has "dos" signature, cns-deploy will never finish due to that. 

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

cns-deploy-3.1.0-14.el7rhgs.x86_64

How reproducible:

always 

Steps to Reproduce:
1. try to create cns cluster using devices with "dos" signature on them 

Actual results:
cns-deploy hangs and never gets back

Expected results:
cns-deploy to finish. 

Inside pod there is run below command to create physical volume :  

# pvcreate --metadatasize=128M --dataalignment=256K /dev/sde 

if this very same command is executed on device with "dos" signature it will report back 

# pvcreate --metadatasize=128M --dataalignment=256K /dev/sde 
WARNING: dos signature detected on /dev/sde at offset 510. Wipe it? [y/n]: 

so it will wait, and will not by default delete "dos" signature. 

solution : 

- either run wipefs -a /dev/sdd 
- pvcreate -fy /dev/sdd 

I think it is fine to use these destructive commands as devices will be specified in topology file and it is responsibility of topology.json creator / owner to ensure it is correct. 

Additional info:
NA

Comment 2 Raghavendra Talur 2017-03-31 16:09:38 UTC
In my opinion, we should not force erase a disk with existing partition, but I agree that the command to error out and exit. Will work on this next release.

Comment 4 Raghavendra Talur 2017-06-12 10:58:20 UTC
We might be able to fix this using the feature in https://github.com/kubernetes/kubernetes/pull/33366 . Should experiment.

Comment 5 Raghavendra Talur 2017-07-25 12:41:24 UTC
https://github.com/heketi/heketi/pull/778

Comment 6 Michael Adam 2017-08-16 15:07:25 UTC
patch merged upstream.

Comment 11 Niels de Vos 2018-05-09 13:36:55 UTC
This bugs has been sent to heketi-devel for further discussion. We may still be able to take this in for cns-3.10 if there is a quick agreement.

http://lists.gluster.org/pipermail/heketi-devel/2018-May/000267.html

Comment 12 Niels de Vos 2018-05-15 10:12:39 UTC
There has been an agreement that a --force option would be useful. When passing this option, the data on the device will be destroyed with 'wipefs --all'.

Most likely the option will be called --destroy-existing-data so that users are warned about the potential danger of using it.

Comment 16 Ashmitha Ambastha 2018-07-10 09:28:43 UTC
I've tried to create a CNS cluster using devices with dos signature on them. cns-deploy doesn't hang and continues with the deployment. 
Hence, moving this bug to verified.

Comment 18 errata-xmlrpc 2018-09-12 09:22:12 UTC
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHEA-2018:2686


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