Greg Farnum
1719db2d73
Hadoop: CephFSInterface cleanup:
...
better error catching on JNI function use;
removed a few unused functions;
now takes advantage of implicit typecasting for clearer code
(no more manually creating j_ints and storing results to make a 1-line
function into 3+ lines; no more if statements returning JNI_TRUE/FALSE);
modified the getDir function as it wasn't very robust
about checking for . and .. entries and the Java code does that;
modified the Java code listPaths as it wasn't necessarily filling
the array and didn't compact it!
2009-08-07 15:05:43 -07:00
Greg Farnum
f3db8c5c2f
Hadoop: Added in initialization checks and
...
limited use of Progressables.
2009-08-07 15:05:43 -07:00
Greg Farnum
83c11b9efe
Hadoop: Changed back to isDirectory as that
...
method was un-deprecated.
2009-08-07 15:05:43 -07:00
Greg Farnum
700c2764e2
Hadoop: Throws IOException -> return null; interface requirement.
...
setReplicationRaw is an old method name, and it doesn't need to be overridden.
getReplication was deprecated and no longer exists.
2009-08-07 15:05:43 -07:00
Yehuda Sadeh
788e5dd9ce
radosgw: static linkage
2009-08-07 11:35:08 -07:00
Yehuda Sadeh
d0f58ad8b2
radosgw: rename symbols, use generic name for rgw
2009-08-07 10:46:29 -07:00
Yehuda Sadeh
a7995b79d2
radosgw: rename files to use the generic name
2009-08-07 10:26:07 -07:00
Yehuda Sadeh
1fb90d367d
Merge commit 'origin/unstable' into rados
...
Conflicts:
src/Makefile.am
2009-08-07 08:50:05 -07:00
Sage Weil
38daf9efae
todo
2009-08-06 16:25:58 -07:00
Sage Weil
c49fd9abbb
debian: fix typo in build script
2009-08-06 16:25:58 -07:00
Greg Farnum
e15c40a7c0
Hadoop: libhadoopcephfs now links against libceph
...
rather than incorporating it; the Java code accounts for this and
loads based on a configuration setting -- no more worries
about java.library.path.
2009-08-06 16:05:30 -07:00
Greg Farnum
360834b652
uClient: now cleans up open files on shutdown.
...
Doesn't crash on a failed assert!
2009-08-06 15:28:00 -07:00
Greg Farnum
b6862d6943
Hadoop: Implements getFileBlockLocations;
...
adds ceph_close to CephFileSystem.
2009-08-06 15:06:19 -07:00
Greg Farnum
b639c218a6
uClient/libceph: initial implementation of get_file_stripe_address
2009-08-06 14:41:54 -07:00
Greg Farnum
f42bb1daaf
Hadoop: Makes use of newly-available replication reporting.
2009-08-05 15:48:23 -07:00
Greg Farnum
9bc42b5cab
uClient/libceph: implement get_file_replication
...
add it to libceph.
2009-08-05 15:48:23 -07:00
Greg Farnum
90b0a6cd11
uClient: Change file data function names for clarity;
...
add function stubs for Hadoop use.
2009-08-05 15:48:23 -07:00
Greg Farnum
290233ea20
Hadoop: time conversion between Ceph and Hadoop;
...
ditch filling in the uid/gid since it's meaningless.
2009-08-05 15:48:23 -07:00
Greg Farnum
2387afb53b
uClient: fill_stat now fills times properly
2009-08-05 15:48:23 -07:00
Sage Weil
f524150822
v0.12
2009-08-05 14:12:00 -07:00
Sage Weil
dad494c563
Makefile: add missing header to noinst_HEADERS
2009-08-05 14:11:17 -07:00
Sage Weil
1b89c167e0
Makefile: remove java hadoop _SOURCES dep
2009-08-05 14:09:59 -07:00
Sage Weil
6b4d8f9fbe
osdmap: kill unused/useless get_num_osds()
2009-08-05 11:41:45 -07:00
Sage Weil
2a1f32944c
mds: show journal entry lengths in debug output
2009-08-05 11:40:05 -07:00
Sage Weil
1fd35e1aac
kclient: cleanup unused label
2009-08-05 11:38:42 -07:00
Sage Weil
f539036d34
kclient: revamp fsync
...
Be smarter about when we write back caps on fsync, and when we
wait. Also, wait only for those caps to write back, not for all
caps to be clean, avoiding starvation.
2009-08-05 11:38:32 -07:00
Sage Weil
f96501a5a1
mds: include mds # in table names
...
Silly us, we were using the same object names for multiple mds's.
That doesn't work so well.
Renamed like so:
mdstable_snaptable -> mds_snaptable
mdstable_inotable -> mds0_inotable
sessionmap -> mds0_sessionmap
mdstable_anchortable -> mds_anchortable
2009-08-04 16:41:36 -07:00
Sage Weil
660ca27162
mds: wait for rejoin_gather_finish to complete before finishing
...
To do that we add ourselves to the rejoin_ack_gather. Otherwise
we end up in up:active before we've even finished our
parallel_fetch or finished up our caps!
2009-08-04 16:08:32 -07:00
Sage Weil
f8501abc72
mds: set primary lock state to LOCK from strong replica when appropriate
...
This is needed only because we identify_files_to_recover() before
sending the rejoin acks, and that may twiddle the lock state, so
we need to be in a compatible state.
2009-08-04 16:07:24 -07:00
Sage Weil
7032e8a084
Makefile: reorg
2009-08-04 15:33:37 -07:00
Sage Weil
80b9480eaa
mds: fix lock state assert on sync->excl
2009-08-04 15:33:36 -07:00
Sage Weil
2eb73630c9
mds: set up root inode properly on replay
2009-08-04 15:33:36 -07:00
Sage Weil
72a1dc3bb8
mds: fix reversed conditional on simplelock ->excl transition
2009-08-04 15:33:36 -07:00
Sage Weil
b3a1b789dd
Makefile: no fakefuse
2009-08-04 15:33:36 -07:00
Sage Weil
55e713b553
psim: calc stddev
2009-08-04 15:33:36 -07:00
Sage Weil
841bc5f885
osdmap: only use domains if ndom > 1
2009-08-04 15:33:36 -07:00
Sage Weil
55f4e003a9
crush: fix grammar so we can recompile what we decompiled
2009-08-04 15:33:36 -07:00
Sage Weil
2c49cc3f25
mds: start resolve with root as UNKNOWN (if it's not ours)
...
Anything that's not ours should be unknown, including the root dir frag,
which normally starts out as mds0.
If we leave it as 0, then when mds0 claims a subset of /, its bounds are
left as 0 as well instead of being set to unknown. Which leads to
incorrect resolve stage results.
2009-08-04 15:33:36 -07:00
Sage Weil
f663a3f04e
mds: fix typo (C_Locker_Eval crash)
2009-08-04 15:33:36 -07:00
Sage Weil
f0b0ce4026
kclient: fix locking snafu in ESTALE handling
2009-08-04 15:33:36 -07:00
Sage Weil
128e9d25c8
kclient: choose mds based on dentry parent, if needed
2009-08-04 15:33:36 -07:00
Sage Weil
40f1ab8efd
mds: adjust mds state asserts to include clientreplay
2009-08-04 15:33:35 -07:00
Sage Weil
da7793a360
mds: twiddle missing locks on export
2009-08-04 15:33:35 -07:00
Sage Weil
0d5ef777a7
mds: open_root only after up:active
...
Otherwise our lock states get totally out of whack.
2009-08-04 15:33:35 -07:00
Sage Weil
84300ab6d1
mds: issue caps on non-auth inodes
...
Not sure how that && is_auth() snuck in there.
2009-08-04 15:33:35 -07:00
Sage Weil
6976f2c1c7
kclient: use caps, fragtree only to choose mds (not hierarchy)
...
Since we require caps for all inodes in our cache, no need to consider
parents when identifying where to sent a request. Just look at fragtree
(for fragmented dirs) or caps.
2009-08-04 15:33:35 -07:00
Sage Weil
849bae88f2
kclient: remove stray ; in dout macro
2009-08-04 15:33:35 -07:00
Greg Farnum
df3be325c1
Hadoop: Replaced some unclean file references.
...
Hey cool, it can do a map-reduce now!
2009-08-04 14:44:48 -07:00
Greg Farnum
74902763b6
uClient: chmod now protects unchangeable bits.
2009-08-04 14:11:59 -07:00
Greg Farnum
7e19b03591
uClient: _rename doesn't remove nodes from cache.
...
Sage says this is unnecessary.
2009-08-04 12:43:34 -07:00
