[Deepsea-users] SES4, What is the correct process to remove cephfs?

Boyd Memmott Boyd.Memmott at suse.com
Fri Jun 9 12:05:40 MDT 2017


I was originally thinking down these lines.  But did not connect the dots with Stage 5.  
Good information.

thanks
Boyd Memmott
SLES L2 Support Engineer
Email: boyd.memmott at suse.com 

-----Original Message-----
From: deepsea-users-bounces at lists.suse.com [mailto:deepsea-users-bounces at lists.suse.com] On Behalf Of Eric Jackson
Sent: Thursday, June 01, 2017 11:08 AM
To: Discussions about the DeepSea management framework for Ceph
Subject: Re: [Deepsea-users] SES4, What is the correct process to remove cephfs?

Hi Boyd,
  Sorry for the delay, but your message was waiting for a moderator.

  Removing the role and adding it back should be sufficient as well. 
The general procedure for removing and re-adding any Ceph functionality is

1) comment out/edit the line in your policy.cfg to exclude the role assignment for the specific node

2) Run Stage 2-5 (Technically, you can run Stage 2 and Stage 5 if you *know* that you are not adding or migrating anything else.)

3) Uncomment/restore the previous policy.cfg

4) Run Stage 2-4 

If you are looking for what steps are really performed, take a look in /srv/salt/ceph/remove and /srv/salt/ceph/rescind.  The remove steps happen on the master node (i.e. typically need Ceph admin keyring access).  The rescind steps happen on the specific node.

Eric

On Thursday, June 01, 2017 06:30:32 PM Jan Fajerski wrote:
> Hi Boyd,
> I came across this recently too and I'll add the functionality to 
> DeepSea before SES5 is released.
> Until then you need to deactivate the MDS daemon(s) before removing 
> the file system. The following steps should get you there:
> 
> salt '*' cmd.run 'systemctl stop ceph-mds.target' # stop all mds 
> daemons # then on the master run:
> ceph mds fail 0 # repeat this for every mds you have # now you can 
> remove the file system with ceph fs rm cephfs –yes-i-really-mean-it # 
> start up your mds daemons again salt '*' cmd.run 'systemctl start 
> ceph-mds.target'
> # and you're read to recreate your CephFS
> 
> Hope that helps!
> Best,
> Jan
> 
> On Tue, May 30, 2017 at 05:51:56PM +0000, Boyd Memmott wrote:
> >   Hi All
> >   
> >   
> >   I am somewhat new to SES product.  I have been experimenting with
> >   cephfs and would like to remove and add again.  But do not find
> >   documentation on the process. I did install ceph with deapsea.  I
> >   commented out the role-mds in policy.cfg and ran “salt-run state.orch
> >   ceph.stage.2” and 3.
> >   
> >   
> >   Yet, “ceph fs rm cephfs –yes-i-really-mean-it” returns “ Error EINVAL:
> >   all MDS deamons must be inactive before removing filesystem”
> >   
> >   
> >   Any suggestions would be appreciated.
> >   
> >   
> >   Thank you
> >   
> >   Boyd Memmott
> >   SLES L2 Support Engineer
> >   
> >   Email: boyd.memmott at suse.com
> >   [1]SUSE
> >
> >References
> >
> >   1. http://www.suse.com/
> >
> >_______________________________________________
> >Deepsea-users mailing list
> >Deepsea-users at lists.suse.com
> >http://lists.suse.com/mailman/listinfo/deepsea-users


More information about the Deepsea-users mailing list