Hello Ceph team,
I have seen some pretty strange behavior of ceph release 0.15 under
Ubuntu 9.0.4 (gcc 4.3.3) and maybe you are interested in it:
First, I have got the following compile error message under Ubuntu
9.0.4 (gcc 4.3.3)
> st_ioctls.c' || echo './'`client/test_ioctls.c
> In function ‘open’,
> inlined from ‘main’ at client/test_ioctls.c:34:
> /usr/include/bits/fcntl2.h:51: error: call to ‘__open_missing_mode’
> declared with attribute error: open with O_CREAT in second argument
> needs 3 arguments
The reason is that the open call in the following code of client/
test_ioctrls.c has no third parameter.
It can easily be fixed by adding e.g. 0x644.
The other problem is this:
Another problem I have normally mounting ceph via cfuse as well as via
the RADOS S3 Gateway is:
> >> cfuse -m <ip of monitor>:6789/ /mnt/ceph/
> bound to 0.0.0.0:6802/5304/0, mounting ceph
> terminate called after throwing an instance of
> 'buffer::end_of_buffer*'
The backtrace is (similar if using rgw or rgw_admin):
> 0x00007f21d2c915ca in __cxa_throw () from /usr/lib/libstdc++.so.6
> #6 0x000000000049e762 in buffer::list::iterator::copy
> (this=0x7f21d0fabed0, len=2, dest=0x7f21d0fabe10 "") at ./include/
> buffer.h:535
> #7 0x0000000000527bbc in decode (m...@0x7fffdb5321f8,
> p...@0x7f21d0fabed0) at ./include/encoding.h:33
> #8 0x00000000005256aa in MonClient::handle_mount_ack
> (this=0x7fffdb5321f0, m=0x25c9970) at mon/MonClient.cc:279
> #9 0x00000000005260dd in MonClient::ms_dispatch
> (this=0x7fffdb5321f0, m=0x25c9970) at mon/MonClient.cc:185
> #10 0x000000000045046b in SimpleMessenger::Endpoint::dispatch_entry
> (this=0x25c2e70) at msg/Messenger.h:100
> #11 0x000000000045670d in
> SimpleMessenger::Endpoint::DispatchThread::entry (this=<value
> optimized out>) at msg/SimpleMessenger.h:303
The mon log file states:
> 09.10.05 10:53:14.434449 mon0(leader).paxosservice(clientmap)
> dispatch client_mount from client? 172.16.217.184:6802/6679/0
> 09.10.05 10:53:14.434469 mon0(leader).client v1 preprocess_query
> client_mount from client? 172.16.217.184:6802/6679/0
> 09.10.05 10:53:14.434500 log [SEC] : client
> 172.16.217.184:6802/6679/0 is not authorized to mount
> 09.10.05 10:53:14.434560 140099797936464 -- 172.16.217.184:6789/0/0
> --> client? 172.16.217.184:6802/6679/0 -- client_mount_ack(client-1
> -1) -- ?+0 0x24b3e10
> 09.10.05 10:53:14.434634 mon0(leader).log v15 logging 09.10.05
> 10:53:14.434507 mon0 172.16.217.184:6789/0/0 16 : [SEC] client
> 172.16.217.184:6802/6679/0 is not authorized to mount
Probably the communication of the message between monitor and client
is not totally correct.
However, the error is strange anyway, because the ceph.conf states:
[mount /]
allow = %everyone
The client is treaded as not authorized, because the ExportEntry of
the mountpoint "/" had no ExportAddrEntry items.
Reason for that: The name of the group in the GroupManager::get_group
method was not trimmed (it was still " %everyone") which fails,
because only the first character is tested to be "%". The was caused
by the cut and pasting from the browser which included some strange
UNICODE fragments that looked like a normal space in various editor
incl. vi, gdb, but it was not and therefore also not trimmed correctly.
> od -c ceph.conf
[ m o u n t / ] \n a l l o w = 302
240 % e v e r y o n e \n
After changing the "space", ceph mounted fine. It would be great if
the configuration parser would reject files with invisible characters.
I don't know if this observations help. Thanks for the good work.
Best Regards
Dirk
--
Dirk Meister | [email protected]
Paderborn Center for Parallel Computing (PC2)
University of Paderborn
Fürstenallee 11 | 33102 Paderborn | Germany
------------------------------------------------------------------------------
Come build with us! The BlackBerry® Developer Conference in SF, CA
is the only developer event you need to attend this year. Jumpstart your
developing skills, take BlackBerry mobile applications to market and stay
ahead of the curve. Join us from November 9-12, 2009. Register now!
http://p.sf.net/sfu/devconf
_______________________________________________
Ceph-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ceph-devel