Sage Weil
142b779f95
cephx: return expires service keys from rotatingkeyring
...
Otherwise there's no point in keeping around old service tickets.
To prevent really old tickets from working, we need to rotate
keys. We want slightly old ones to still work, though.. that's
why we keep 3.
2010-02-11 10:03:25 -08:00
Sage Weil
8cbfb327a8
cephx: fix negotiation on reconnect
...
Don't send another request after initial handshake if we don't
need an auth ticket.
2010-02-11 10:03:25 -08:00
Sage Weil
44dc114385
monclient: renew service tickets a bit after the current expires
...
This ensures the monitor will actually ahve generated a newer
one to give us, avoiding a busy loop.
2010-02-11 10:03:25 -08:00
Yehuda Sadeh
ef54e0ccd0
osd: don't update object size if didn't write anything
2010-02-10 17:19:08 -08:00
Sage Weil
2feea3cb56
cmon: suggest mkcephfs when 'whoami' not in monfs
2010-02-10 15:00:57 -08:00
Sage Weil
2f900189bd
cephx: fix up key rotation
2010-02-10 15:00:57 -08:00
Sage Weil
9be2e6a58d
cephx: nicer keyserver debug output
2010-02-10 15:00:56 -08:00
Yehuda Sadeh
e401cd242b
osd: write op updates trancation sequence if not already set
2010-02-10 14:36:10 -08:00
Greg Farnum
3fb7364312
msgr: Update 'documentation'.
2010-02-10 12:02:56 -08:00
Sage Weil
878f5fc388
init-ceph, mkcephfs: fix ETCDIR
2010-02-10 11:33:23 -08:00
Sage Weil
9662ad40e0
mds: behave when we pipeline session updates to journal
2010-02-10 10:31:16 -08:00
Greg Farnum
159dc008f6
msg: union sockaddr_storage to hush strict aliasing warnings and clean up code
2010-02-09 16:36:57 -08:00
Sage Weil
53d3df0ef8
init-ceph: Required-start: $remote_fs
...
This ensures /usr is mounted before ceph daemons start. It seems like
this may be problematic for hosts that act as both servers and clients,
but nfs-kernel-server does the same, so whatev!
2010-02-09 12:28:25 -08:00
Sage Weil
469d089bec
debian: do not include var/run/ceph in package; mkdir -p dirname unconditionally
2010-02-09 12:28:25 -08:00
Sage Weil
2d745bc517
rados man page; include rados in ceph package
2010-02-09 12:28:14 -08:00
Sage Weil
789d702687
debian: include cauthtool
2010-02-09 12:28:07 -08:00
Sage Weil
c77dd333c5
rename authtool -> cauthtool
2010-02-09 10:14:04 -08:00
Sage Weil
d6d3d2ab06
mds: behave when we pipeline session updates to journal
2010-02-09 10:03:12 -08:00
Sage Weil
ad9f8dbc88
workqueue: make drain() wait for processing AND queue
2010-02-09 10:02:41 -08:00
Sage Weil
c77b536b07
mutex: fix file mode
2010-02-09 09:58:54 -08:00
Sage Weil
6aea5ccbcf
osd: take ondisk_read_lock if pushing object to peer
2010-02-09 09:03:51 -08:00
Sage Weil
a3a248cb31
filestore: return 0 for no-op transaction
2010-02-09 08:35:33 -08:00
Sage Weil
b06b232540
debian: fixups to build inside pbuilder
2010-02-09 08:24:57 -08:00
Josef Bacik
270a7bdd94
ceph: fix manpages so they are only installed once
...
While creating a spec file for CEPH, rpmbuild was complaining because make
install was copying the manpages in, and then copying them in again. This is
because man_MANS and dist_man_MANS are supposed to be two seperate lists that do
not overlap. So make install would install all the man pages in the man_MANS
list and the dist_man_MANS list. This patch kills the dist_man_MANS thing to
keep this from happening. This made rpmbuild happy, which makes me happy :).
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
2010-02-09 08:24:41 -08:00
Sage Weil
3f429e005e
Merge branch 'unstable' into filestore
2010-02-09 08:15:31 -08:00
Sage Weil
0ef0ab4b45
osd: count objects degraded due to degraded pg
2010-02-09 08:15:20 -08:00
Sage Weil
62a0e4bb98
osd: prevent do_waiters() while _dispatch() is running
...
Fixes race between handle_osd_map and tick() requeuing ops.
2010-02-09 08:15:20 -08:00
Sage Weil
4a8d1ecdbf
ceph: wait for monmap
2010-02-09 08:15:20 -08:00
Sage Weil
03da8ba906
osd: btrfs_snap by default
2010-02-09 08:09:37 -08:00
Sage Weil
48e3b00f42
osd: carry ondisk_read_lock if prepare_transaction may_read()
2010-02-09 08:09:23 -08:00
Sage Weil
153a37e7ac
osd: order rmdir in remove_collection after deleting dir contents
2010-02-09 08:09:21 -08:00
Sage Weil
816ab37f52
filestore: fix old clobbering of commit_op_seq
...
Update only at commit. And add a newline.
2010-02-09 08:09:19 -08:00
Sage Weil
408c81c210
osd: fix lock inversion with ondisk_write_unlock
...
Drop write lock before taking pg lock.
2010-02-09 08:09:16 -08:00
Sage Weil
c91716edee
osd: store local osd magic, whoami, and other static bits outside of ObjectStore
...
These values are immutable, and we also want to look at them prior to
forking and 'mounting' the ObjectStore. Just keep them in separate files
for simplicity.
This avoids the double filestore startup cost paid on cosd startup.
2010-02-08 20:29:09 -08:00
Sage Weil
73cc468312
osd: print truncate_size signed, and only print at all if _seq > 0
2010-02-08 12:47:17 -08:00
Sage Weil
0f42df4685
debian: updated debian build scripts, changelog
2010-02-08 10:27:49 -08:00
Sage Weil
d87f66d79c
mkcephfs: warn on missing keyring for mds, osd
2010-02-08 10:04:18 -08:00
Sage Weil
4b3c3d31fe
authtool: add -a/--add-key command
2010-02-08 10:04:16 -08:00
Sage Weil
a23cdb0e1c
buffer: add decode_base64 method
2010-02-08 10:04:14 -08:00
Sage Weil
dbcb325e68
cephx: pipe down about ticket renewals
2010-02-08 10:04:11 -08:00
Sage Weil
fdbd893f6a
osd, mds: don't time out authenticate()
...
Still need to fix wait_auth_rotating....
2010-02-06 11:29:39 -08:00
Sage Weil
2ec2d579cb
filejournal: make io contiguous in write_bl() for directio
...
Previously we were splitting the io for writing the header plus first
segment following a wrap.
2010-02-06 11:29:13 -08:00
Sage Weil
d289ea0e62
ceph: error out on authentication failure
...
asdf
2010-02-06 11:29:09 -08:00
Sage Weil
82afbb6003
monc: fix authentication timeout
2010-02-06 11:27:41 -08:00
Sage Weil
26eb29d39f
monclient: kill unused wait_authenticate()
2010-02-06 11:18:38 -08:00
Sage Weil
29d5dd671d
objectstore: include struct_v for Transaction
2010-02-06 10:39:21 -08:00
Sage Weil
ffb1b6c0b0
osd: wait for readable on primary after pulling an object
...
Otherwise our attempt to push it back out again fail.
Note that currently this disables the previous push-right-after-
pull behavior. We should either optimize the common case (full,
non-snapped object), or at least requeue the object so it's at
the front of the recovery queue so the fs cache will be hot.
2010-02-05 16:38:41 -08:00
Sage Weil
f668ce95ab
authtool: only create keyring if --create-keyring (or -c)
2010-02-05 14:16:48 -08:00
Sage Weil
0a6e4ed38a
mkcephfs: misc cephx fixes
2010-02-05 14:11:49 -08:00
Sage Weil
c76ea9de70
debian: fix builder; add newish headers to Makefile.am
2010-02-05 14:11:39 -08:00