On 11/08/14 20:52, Mark Kirkwood wrote:
On 07/08/14 11:06, Mark Kirkwood wrote:
Hi,
I'm doing a fresh install of ceph 0.83 (src build) to an Ubuntu 14.04 VM
using ceph-deploy 1.59. Everything goes well until the osd creation,
which fails to start with a journal open error. The steps are shown
below (ceph is the deploy target host):
$ tail ceph.osd.0.log
2014-08-07 10:47:45.350623 7ffe95e05800 1 journal _open
/var/lib/ceph/osd/ceph-0/journal fd 20: 2147483648 bytes, block size
4096 bytes, directio = 1, aio = 1
2014-08-07 10:47:45.351364 7ffe95e05800 -1 journal read_header error
decoding journal header
2014-08-07 10:47:45.351398 7ffe95e05800 -1
filestore(/var/lib/ceph/osd/ceph-0) mount failed to open journal
/var/lib/ceph/osd/ceph-0/journal: (22) Invalid argument
Doing some more analysis pointed to something fishy with separate
journals. In an effort to get to the root cause I decided create a ceph
setup directly via a simple script (attached) and so avoid issues about
whether ceph-deploy was a factor.
It quickly emerged that the issue was todo with a recent commit
concerning journals. In a simplified test case where I'm trying to
create a single osd with a separate device (or device partition) for a
journal I'm seeing hangs or 'invalid argument 22' errors. Using ceph
version 0.83-611-g4d2d4dd if I revert commit 4eb18dd I find that I can
suddenly get osds up in a fresh install [1], and in upgrades I can
rescue osd's with device journals that refuse to start *provided* I
recreate the journal (yes that's a little strange, possibly some more
commits to examine...but at least I can get 'em started)!
regards
Mark
[1] Specifically in fresh install hanging on a mutex at osd mkfs step:
futex(0x7fffaa3fcbac, FUTEX_WAIT_PRIVATE, 1, NULL...)
and in existing setups seeing:
filestore(/var/lib/ceph/osd/ceph-0) mount failed to open journal
/var/lib/ceph/osd/ceph-0/journal: (22) Invalid argument
as indicated above.
...and this looks like it is something better raised on the dev list or
as a bug (will do the latter).
Mark
_______________________________________________
ceph-users mailing list
[email protected]
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
