Re: [ceph-users] Reinstall everything

2018-06-11 Thread Max Cuttins

I used ceph-deploy to purge the data.
However it didn't remove LVM.

I'm gonna to see the "lvm" part.
Seems new, i used always disk instead since now.




Il 10/06/2018 14:13, Sergey Malinin ha scritto:
Not sure if ceph-deploy has similar functionality, but executing 
‘ceph-volume lvm zap  --destroy’ on target machine would have 
removed lvm mapping.

On Jun 10, 2018, 14:41 +0300, Max Cuttins , wrote:


I solved by myself.
I' writing here my findings to save some working hours to others.
Sound strange that nobody knew this.

The issue is that data is purged but LVM partition are leaved in place.

This means that you need to manually remove.

I just reinstalled the whole OS and on the data disks there are still 
LVM partition named "ceph-*". These partition are ACTIVE by default.

To get rid of the old data:

#find disks

lsblk

See in the result all the "ceph-*" volume groups and remove theme:

vgchange -a n ceph-XX
vgremove ceph-XXX

Do it for all disks.
Now you can run *ceph-deploy osd create* correctly without being 
prompted that disk is in use.





Il 06/06/2018 19:41, Max Cuttins ha scritto:

Hi everybody,

I would like to start from zero.
However last time I run the command to purge everything I got an issue.

I had a complete cleaned up system as expected, but disk was still 
OSD and the new installation refused to overwrite disk in use.
The only way to make it work was manually format the disks with 
fdisk and zap again with ceph later.


Is there something I shoulded do before purge everything in order to 
do not have similar issue?


Thanks,
Max
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com



___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com


___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com


Re: [ceph-users] Reinstall everything

2018-06-10 Thread Sergey Malinin
Not sure if ceph-deploy has similar functionality, but executing ‘ceph-volume 
lvm zap  --destroy’ on target machine would have removed lvm mapping.
On Jun 10, 2018, 14:41 +0300, Max Cuttins , wrote:
> I solved by myself.
> I' writing here my findings to save some working hours to others.
> Sound strange that nobody knew this.
> The issue is that data is purged but LVM partition are leaved in place.
> This means that you need to manually remove.
> I just reinstalled the whole OS and on the data disks there are still LVM 
> partition named "ceph-*". These partition are ACTIVE by default.
> To get rid of the old data:
> #find disks
> > lsblk
> See in the result all the "ceph-*" volume groups and remove theme:
> > vgchange -a n ceph-XX
> > vgremove ceph-XXX
> Do it for all disks.
> Now you can run ceph-deploy osd create correctly without being prompted that 
> disk is in use.
>
>
>
> Il 06/06/2018 19:41, Max Cuttins ha scritto:
> > Hi everybody,
> >
> > I would like to start from zero.
> > However last time I run the command to purge everything I got an issue.
> >
> > I had a complete cleaned up system as expected, but disk was still OSD and 
> > the new installation refused to overwrite disk in use.
> > The only way to make it work was manually format the disks with fdisk and 
> > zap again with ceph later.
> >
> > Is there something I shoulded do before purge everything in order to do not 
> > have similar issue?
> >
> > Thanks,
> > Max
> > ___
> > ceph-users mailing list
> > ceph-users@lists.ceph.com
> > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
> >
>
> ___
> ceph-users mailing list
> ceph-users@lists.ceph.com
> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com


Re: [ceph-users] Reinstall everything

2018-06-10 Thread Max Cuttins

I solved by myself.
I' writing here my findings to save some working hours to others.
Sound strange that nobody knew this.

The issue is that data is purged but LVM partition are leaved in place.

This means that you need to manually remove.

I just reinstalled the whole OS and on the data disks there are still 
LVM partition named "ceph-*". These partition are ACTIVE by default.

To get rid of the old data:

#find disks

   lsblk

See in the result all the "ceph-*" volume groups and remove theme:

   vgchange -a n ceph-XX
   vgremove ceph-XXX

Do it for all disks.
Now you can run *ceph-deploy osd create* correctly without being 
prompted that disk is in use.





Il 06/06/2018 19:41, Max Cuttins ha scritto:

Hi everybody,

I would like to start from zero.
However last time I run the command to purge everything I got an issue.

I had a complete cleaned up system as expected, but disk was still OSD 
and the new installation refused to overwrite disk in use.
The only way to make it work was manually format the disks with fdisk 
and zap again with ceph later.


Is there something I shoulded do before purge everything in order to 
do not have similar issue?


Thanks,
Max
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com



___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com


[ceph-users] Reinstall everything

2018-06-06 Thread Max Cuttins

Hi everybody,

I would like to start from zero.
However last time I run the command to purge everything I got an issue.

I had a complete cleaned up system as expected, but disk was still OSD 
and the new installation refused to overwrite disk in use.
The only way to make it work was manually format the disks with fdisk 
and zap again with ceph later.


Is there something I shoulded do before purge everything in order to do 
not have similar issue?


Thanks,
Max
___
ceph-users mailing list
ceph-users@lists.ceph.com
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com