All of lore.kernel.org
 help / color / mirror / Atom feed
* mkcephfs error: cannot remove `/var/lib/ceph/osd/ceph-0'
@ 2013-02-09  3:39 Adam Nielsen
  2013-02-09  4:55 ` Sage Weil
  0 siblings, 1 reply; 4+ messages in thread
From: Adam Nielsen @ 2013-02-09  3:39 UTC (permalink / raw)
  To: ceph-devel

Hi all,

I'm trying to set up a test Ceph node, but mkcephfs is failing because it's 
trying to erase my mount point, which it can't do because the disk I want to 
use with Ceph is mounted there.  I finally found some manual installation 
steps in the mkcephfs manpage (I think these really should be included in the 
online docs!) and this is the bit that fails:

$ mkcephfs --init-local-daemons mon -d /tmp/foo
=== mon.0 ===
rm: cannot remove `/var/lib/ceph/osd/ceph-0': Device or resource busy
/usr/bin/ceph-mon: error creating monfs: (5) Input/output error
2013-02-09 13:17:24.941676 b7358710 -1 store(/var/lib/ceph/osd/ceph-0/) 
MonitorStore::mkfs: failed to remove /var/lib/ceph/osd/ceph-0/: rm returned 
run_cmd(rm): exited with status 1
2013-02-09 13:17:24.941746 b7358710 -1 mon.0@-1(probing) e0 store->mkfs failed 
with: (5) Input/output error

What am I doing wrong?  I am going off the ceph.com docs[1], which says "If 
you are running multiple OSDs per host and one hard disk per OSD, you should 
mount the disk under the OSD data directory (if you haven’t done so already)". 
  I am only running one OSD on this host, does this mean I shouldn't mount the 
disk before running mkcephfs?  Won't that just create all the files on the 
wrong disk?

I'm running Debian Squeeze with the debian-bobtail packages installed as per 
the ceph.com Debian instructions[2].

Any suggestions?

Many thanks,
Adam.

[1] http://ceph.com/docs/master/rados/deployment/mkcephfs/
[2] http://ceph.com/docs/master/install/debian/

P.S. I would've posted on ceph-users except it's not carried on GMane yet.

--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: mkcephfs error: cannot remove `/var/lib/ceph/osd/ceph-0'
  2013-02-09  3:39 mkcephfs error: cannot remove `/var/lib/ceph/osd/ceph-0' Adam Nielsen
@ 2013-02-09  4:55 ` Sage Weil
  2013-02-09  5:39   ` Adam Nielsen
  0 siblings, 1 reply; 4+ messages in thread
From: Sage Weil @ 2013-02-09  4:55 UTC (permalink / raw)
  To: Adam Nielsen; +Cc: ceph-devel

On Sat, 9 Feb 2013, Adam Nielsen wrote:
> Hi all,
> 
> I'm trying to set up a test Ceph node, but mkcephfs is failing because it's
> trying to erase my mount point, which it can't do because the disk I want to
> use with Ceph is mounted there.  I finally found some manual installation
> steps in the mkcephfs manpage (I think these really should be included in the
> online docs!) and this is the bit that fails:
> 
> $ mkcephfs --init-local-daemons mon -d /tmp/foo

Not sure if this is the problem, but: you're creating the mon...

> === mon.0 ===
> rm: cannot remove `/var/lib/ceph/osd/ceph-0': Device or resource busy

..but the path is /var/lib/ceph/osd/...?  Did you override the paths in 
your ceph.conf?

sage

> /usr/bin/ceph-mon: error creating monfs: (5) Input/output error
> 2013-02-09 13:17:24.941676 b7358710 -1 store(/var/lib/ceph/osd/ceph-0/)
> MonitorStore::mkfs: failed to remove /var/lib/ceph/osd/ceph-0/: rm returned
> run_cmd(rm): exited with status 1
> 2013-02-09 13:17:24.941746 b7358710 -1 mon.0@-1(probing) e0 store->mkfs failed
> with: (5) Input/output error
> 
> What am I doing wrong?  I am going off the ceph.com docs[1], which says "If
> you are running multiple OSDs per host and one hard disk per OSD, you should
> mount the disk under the OSD data directory (if you haven?t done so already)".
> I am only running one OSD on this host, does this mean I shouldn't mount the
> disk before running mkcephfs?  Won't that just create all the files on the
> wrong disk?
> 
> I'm running Debian Squeeze with the debian-bobtail packages installed as per
> the ceph.com Debian instructions[2].
> 
> Any suggestions?
> 
> Many thanks,
> Adam.
> 
> [1] http://ceph.com/docs/master/rados/deployment/mkcephfs/
> [2] http://ceph.com/docs/master/install/debian/
> 
> P.S. I would've posted on ceph-users except it's not carried on GMane yet.
> 
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: mkcephfs error: cannot remove `/var/lib/ceph/osd/ceph-0'
  2013-02-09  4:55 ` Sage Weil
@ 2013-02-09  5:39   ` Adam Nielsen
  2013-02-09  5:42     ` Sage Weil
  0 siblings, 1 reply; 4+ messages in thread
From: Adam Nielsen @ 2013-02-09  5:39 UTC (permalink / raw)
  To: ceph-devel

>> $ mkcephfs --init-local-daemons mon -d /tmp/foo
>
> Not sure if this is the problem, but: you're creating the mon...
>
>> === mon.0 ===
>> rm: cannot remove `/var/lib/ceph/osd/ceph-0': Device or resource busy
>
> ..but the path is /var/lib/ceph/osd/...?  Did you override the paths in
> your ceph.conf?

Ah yes I did, I see the problem now.  What I wanted to do was to put *all* 
Ceph data onto the same disk (since I am booting the machine off an SD card so 
there is not enough space for much else.)

What's the best solution in this case?  Should I override the mon path and put 
it as /var/lib/ceph/osd/ceph-0/mon/ for example, or should I mount my disk as 
/var/lib/ceph instead?

Many thanks,
Adam.



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: mkcephfs error: cannot remove `/var/lib/ceph/osd/ceph-0'
  2013-02-09  5:39   ` Adam Nielsen
@ 2013-02-09  5:42     ` Sage Weil
  0 siblings, 0 replies; 4+ messages in thread
From: Sage Weil @ 2013-02-09  5:42 UTC (permalink / raw)
  To: Adam Nielsen; +Cc: ceph-devel

On Sat, 9 Feb 2013, Adam Nielsen wrote:
> > > $ mkcephfs --init-local-daemons mon -d /tmp/foo
> > 
> > Not sure if this is the problem, but: you're creating the mon...
> > 
> > > === mon.0 ===
> > > rm: cannot remove `/var/lib/ceph/osd/ceph-0': Device or resource busy
> > 
> > ..but the path is /var/lib/ceph/osd/...?  Did you override the paths in
> > your ceph.conf?
> 
> Ah yes I did, I see the problem now.  What I wanted to do was to put *all*
> Ceph data onto the same disk (since I am booting the machine off an SD card so
> there is not enough space for much else.)
> 
> What's the best solution in this case?  Should I override the mon path and put
> it as /var/lib/ceph/osd/ceph-0/mon/ for example, or should I mount my disk as
> /var/lib/ceph instead?

Yeah, you could mount it on /var/lib/ceph, or just make a symlink.

sage

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2013-02-09  5:42 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-02-09  3:39 mkcephfs error: cannot remove `/var/lib/ceph/osd/ceph-0' Adam Nielsen
2013-02-09  4:55 ` Sage Weil
2013-02-09  5:39   ` Adam Nielsen
2013-02-09  5:42     ` Sage Weil

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.