Bug 1388557 - multipath-tools: rbd checker should not assume default cluster name
Summary: multipath-tools: rbd checker should not assume default cluster name
Keywords:
Status: CLOSED WONTFIX
Alias: None
Product: Red Hat Enterprise Linux 7
Classification: Red Hat
Component: device-mapper-multipath
Version: 7.4
Hardware: Unspecified
OS: Unspecified
medium
unspecified
Target Milestone: rc
: ---
Assignee: Mike Christie
QA Contact: Lin Li
URL:
Whiteboard:
Depends On:
Blocks: 1379890
TreeView+ depends on / blocked
 
Reported: 2016-10-25 16:04 UTC by Mike Christie
Modified: 2021-09-03 12:09 UTC (History)
8 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2017-01-04 21:39:10 UTC
Target Upstream Version:


Attachments (Terms of Use)

Description Mike Christie 2016-10-25 16:04:08 UTC
Description of problem:

If the ceph cluster was not using the default ceph name, then rbd checker initialization will fail with

Oct 25 04:52:19 | rbd2: Could not read rados conf

like is seen in 

https://bugzilla.redhat.com/show_bug.cgi?id=1388063#c4

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


How reproducible:


Steps to Reproduce:
1.
2.
3.

Actual results:


Expected results:


Additional info:

Comment 3 Tejas 2016-10-28 09:08:41 UTC
Mike,

   Customers will have ceph clusters with different names. We cannot ask them to recreate ceph cluster for the iscsi gateways. So this is  a critical fix from the view of backward compatibility.
Any idea when we can get this fix? We need this fix.

Thanks,
Tejas

Comment 4 Mike Christie 2016-10-28 16:49:15 UTC
(In reply to Tejas from comment #3)
> Mike,
> 
>    Customers will have ceph clusters with different names. We cannot ask
> them to recreate ceph cluster for the iscsi gateways. So this is  a critical
> fix from the view of backward compatibility.
> Any idea when we can get this fix? We need this fix.
> 


I should hopefully have something upstream approves of Monday or Tuesday. I can make you a rpm then.

Comment 7 Jason Dillaman 2016-11-02 16:06:47 UTC
I flagged the ceph-iscsi-ansible side of this issue (BZ 1388063) to be included in the release notes. You don't need to re-create your cluster. The cluster name is just a way to automatically pick a configuration file. Instead, if you gave your cluster a unique name, you just need to copy your "/etc/ceph/<unique cluster name>.conf" (and associated keyring) to "/etc/ceph/ceph.conf" on the bootstrap monitor host.

Comment 9 Mike Christie 2017-01-04 21:39:10 UTC
We are changing the implementation in RHCS 2.2 so we will not need this change anymore.


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