Re: bobtail release candidates

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 11/26/2012 08:36 AM, Wido den Hollander wrote:


On 11/26/2012 10:26 PM, Sam Lang wrote:
On 11/26/2012 07:47 AM, Wido den Hollander wrote:
Hi,

On 11/26/2012 01:57 AM, Sage Weil wrote:
Hi all,

There are automatic builds of the prerelease bobtail code available
under
the 'next' branch.

For debs,

http://ceph.com/docs/master/install/debian/#add-development-testing-packages




For example, for Ubuntu 12.04 precise,
    http://gitbuilder.ceph.com/ceph-deb-precise-x86_64-basic/ref/next/

And RPMs for el6,

http://gitbuilder.ceph.com/ceph-rpm-centos6-x86_64-basic/ref/next/RPMS/x86_64/




Any testing and early feedback is greatly appreciated.


I found another issue which I'm not sure about.

I did the upgrade from 0.48.2 to 0.54 by simply running apt-get upgrade.

The mon restarted fine, but when trying to connect I get:

"-1 unable to authenticate as client.admin"

I ran with debug ms/auth = 20 and I got:

2012-11-26 14:41:13.039615 7fed35bc7780  1 -- :/0 messenger.start
2012-11-26 14:41:13.039878 7fed35bc7780 10 -- :/29176 ready :/29176
2012-11-26 14:41:13.039905 7fed32a70700 10 -- :/29176 reaper_entry start
2012-11-26 14:41:13.040309 7fed32a70700 10 -- :/29176 reaper
2012-11-26 14:41:13.040324 7fed32a70700 10 -- :/29176 reaper done
2012-11-26 14:41:13.041255 7fed35bc7780  2 auth: KeyRing::load: loaded
key file /etc/ceph/ceph.keyring
2012-11-26 14:41:13.041943 7fed35bc7780 10 -- :/29176 connect_rank to
192.168.6.250:6789/0, creating pipe and registering
2012-11-26 14:41:13.042119 7fed35bc7780 10 -- :/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=-1 :0 pgs=0 cs=0
l=1).register_pipe
2012-11-26 14:41:13.042329 7fed35bc7780 10 -- :/29176 get_connection
mon.0 192.168.6.250:6789/0 new 0x1361670
2012-11-26 14:41:13.042152 7fed35bc3700 10 -- :/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=-1 :0 pgs=0 cs=0 l=1).writer:
state = connecting policy.server=0
2012-11-26 14:41:13.042408 7fed35bc3700 10 -- :/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=-1 :0 pgs=0 cs=0 l=1).connect 0
2012-11-26 14:41:13.042534 7fed35bc3700 10 -- :/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :0 pgs=0 cs=0 l=1).connecting
to 192.168.6.250:6789/0
2012-11-26 14:41:13.042638 7fed35bc7780  1 -- :/29176 -->
192.168.6.250:6789/0 -- auth(proto 0 26 bytes epoch 0) v1 -- ?+0
0x1361df0 con 0x13618b0
2012-11-26 14:41:13.042690 7fed35bc7780 20 -- :/29176 submit_message
auth(proto 0 26 bytes epoch 0) v1 remote, 192.168.6.250:6789/0, have
pipe.
2012-11-26 14:41:13.043255 7fed35bc3700 20 -- :/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=0 cs=0 l=1).connect
read peer addr 192.168.6.250:6789/0 on socket 3
2012-11-26 14:41:13.043331 7fed35bc3700 20 -- :/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=0 cs=0 l=1).connect
peer addr for me is 192.168.6.250:41567/0
2012-11-26 14:41:13.043403 7fed35bc3700  1 -- 192.168.6.250:0/29176
learned my addr 192.168.6.250:0/29176
2012-11-26 14:41:13.043548 7fed35bc3700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=0 cs=0 l=1).connect
sent my addr 192.168.6.250:0/29176
2012-11-26 14:41:13.043634 7fed35bc3700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=0 cs=0 l=1).connect
sending gseq=1 cseq=0 proto=15
2012-11-26 14:41:13.043723 7fed35bc3700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=0 cs=0 l=1).connect
wrote (self +) cseq, waiting for reply
2012-11-26 14:41:13.043942 7fed35bc3700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=0 cs=0 l=1).connect
got reply tag 1 connect_seq 1 global_seq 5 proto 15 flags 1
2012-11-26 14:41:13.044008 7fed35bc3700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).connect
success 1, lossy = 1, features 33554431
2012-11-26 14:41:13.044147 7fed35bc3700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).connect
starting reader
2012-11-26 14:41:13.044425 7fed35bc3700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).writer:
state = open policy.server=0
2012-11-26 14:41:13.044500 7fed35bc3700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).writer
encoding 1 0x1361df0 auth(proto 0 26 bytes epoch 0) v1
2012-11-26 14:41:13.044740 7fed35bc3700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).writer
sending 1 0x1361df0
2012-11-26 14:41:13.044521 7fed30a6c700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
reading tag...
2012-11-26 14:41:13.044799 7fed35bc3700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).Pipe:
write_message:  session security NULL for this pipe.
2012-11-26 14:41:13.044842 7fed35bc3700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1
l=1).write_message 0x1361df0
2012-11-26 14:41:13.045001 7fed35bc3700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).writer:
state = open policy.server=0
2012-11-26 14:41:13.045048 7fed35bc3700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).writer
sleeping
2012-11-26 14:41:13.045779 7fed30a6c700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
got ACK
2012-11-26 14:41:13.045864 7fed30a6c700 15 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
got ack seq 1
2012-11-26 14:41:13.045906 7fed30a6c700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
reading tag...
2012-11-26 14:41:13.045957 7fed30a6c700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
got MSG
2012-11-26 14:41:13.046015 7fed30a6c700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
got envelope type=18 src mon.0 front=24 data=0 off 0
2012-11-26 14:41:13.046099 7fed30a6c700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
wants 24 from dispatch throttler 0/104857600
2012-11-26 14:41:13.046176 7fed30a6c700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
got front 24
2012-11-26 14:41:13.046230 7fed30a6c700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1
l=1).aborted = 0
2012-11-26 14:41:13.046269 7fed30a6c700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
got 24 + 0 + 0 byte message
2012-11-26 14:41:13.046362 7fed30a6c700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).No
session security set
2012-11-26 14:41:13.046426 7fed30a6c700 10 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
got message 1 0x7fed14001660 auth_reply(proto 0 -95 Operation not
supported) v1
2012-11-26 14:41:13.046536 7fed30a6c700 20 -- 192.168.6.250:0/29176
queue 0x7fed14001660 prio 196
2012-11-26 14:41:13.046595 7fed30a6c700 20 -- 192.168.6.250:0/29176 >>
192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
reading tag...
2012-11-26 14:41:13.046625 7fed2226f700  1 -- 192.168.6.250:0/29176 <==
mon.0 192.168.6.250:6789/0 1 ==== auth_reply(proto 0 -95 Operation not
supported) v1 ==== 24+0+0 (3632875112 0 0) 0x7fed14001660 con 0x13618b0
2012-11-26 14:41:13.046858 7fed2226f700 10 -- 192.168.6.250:0/29176
dispatch_throttle_release 24 to dispatch throttler 24/104857600
2012-11-26 14:41:13.046883 7fed35bc7780 -1 unable to authenticate as
client.admin
2012-11-26 14:41:13.046890 7fed2226f700 20 -- 192.168.6.250:0/29176 done
calling dispatch on 0x7fed14001660
2012-11-26 14:41:13.046904 7fed35bc7780 10 -- 192.168.6.250:0/29176
shutdown 192.168.6

It seems to read /etc/ceph/ceph.keyring, but then it can't connect?

192.168.6.250:6789/0 pipe(0x1361670 sd=3 :41567 pgs=5 cs=1 l=1).reader
got message 1 0x7fed14001660 auth_reply(proto 0 -95 Operation not
supported) v1

Not sure what that means?

My ceph.conf is minimal as possible, there is not even an [global]
section and my mon section only lists one monitor, no other options.

I tried to keep everything to its defaults to keep the setup simple.

Suggestions?


Hi Wido,

The defaults for auth parameters changed in bobtail.  With no settings
in your config, you probably weren't using auth, but after the upgrade,
the defaults became cephx.

I think you can fix it by disabling auth (which is what you had in 0.48):

[global]
auth cluster required = none
auth service required = none


Here's the commit where the options changed:

https://github.com/ceph/ceph/commit/66bda162e1acad34d37fa97e3a91e277df174f42



Argh, should have spotted this one. I thought this had changed in 0.48.2
already, but that was my mistake.

I had to add "auth client required = none" on my client though, since
that still tried cephx.

This is going to trip up a lot of people. On an upgrade from 0.48 to 0.54, can we check that the config is setup this way and either warn the user or add the right config options on their behalf?

-sam


Works now!

Thanks,

Wido


-sam

Wido





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

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


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


[Index of Archives]     [CEPH Users]     [Ceph Large]     [Information on CEPH]     [Linux BTRFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]
  Powered by Linux