Sage Weil
b0c6a3a879
mds: fix stat mask
2009-01-06 12:46:00 -08:00
Sage Weil
1cf4b7145b
mds: only send clientcaps if we need to revoke, or we can issue explicitly wanted caps
2009-01-06 12:45:49 -08:00
Sage Weil
c1d9372cbb
mds: fix cap release seq check
2009-01-06 12:45:24 -08:00
Sage Weil
12f429d640
uclient: set dirty bits in cap_snap
2009-01-06 11:08:44 -08:00
Sage Weil
4a499ee25c
mds: bunch of cap protocol changes, and user client adjustments.
...
See capability protocol notes in mds/Capability.h
2009-01-05 16:30:10 -08:00
Sage Weil
437226eb54
kclient: fix snap realm update on reply
...
Take snap_rwsem always, since we usually get caps. Drop trailing
snaprealm ref after trace is fully processed.
2009-01-05 11:20:13 -08:00
Sage Weil
7ba06acfe3
mds: fix set_trace_dist snaprealm problem on unlink
...
Set snaprealm trace based on the first inode we see.
2009-01-05 11:06:26 -08:00
Sage Weil
886091eff1
kclient: ref count cap_snap to avoid unnecessarily copying into temp variables
2009-01-04 15:41:03 -08:00
Sage Weil
be1d5561b2
kclient: take fmode ref atomically while adding cap
2009-01-04 12:14:40 -08:00
Sage Weil
94e3abfe28
client: do not update when already holding EXCL cap
2009-01-04 11:38:31 -08:00
Sage Weil
fb882d34c1
kclient: caps basically working
...
handle caps on any inode in reply. no more inode leases.
2009-01-04 11:17:20 -08:00
Yehuda Sadeh
f920808ae8
kclient: use ceph_caps_release instead of dropping inode lease
2009-01-04 10:51:29 -08:00
Yehuda Sadeh
f17fabe7fa
kclient: switch from inode leases to new caps
...
----------------------------
2009-01-04 10:50:56 -08:00
Sage Weil
a46dbfd5e8
notes
2009-01-04 10:34:47 -08:00
Sage Weil
6a71cda024
uclient: adjust cap release logic...
2009-01-04 10:28:50 -08:00
Sage Weil
2dfa4db374
mds: only _do_cap_update if client held ANY_WR caps
2009-01-04 09:38:46 -08:00
Sage Weil
8276576ce2
mds: issue RDCACHE caps to client when possible. add cap to do so if inode auth.
2009-01-03 11:33:35 -08:00
Sage Weil
01f28e5b24
uclient: sorta working
2009-01-02 17:03:47 -08:00
Sage Weil
dad9bf47b7
mds: fix scatterlock print
2009-01-02 17:00:08 -08:00
Sage Weil
3c571d46fb
mds: put a cap in each inode stat in trace; simplify snaprealm trace encoding
2009-01-02 15:29:29 -08:00
Sage Weil
e3e3f2d1f7
mds: cap comments
2009-01-02 14:46:07 -08:00
Sage Weil
b4561ba15f
mds: rewrite CEPH_STAT_MASK_* as CEPH_STAT_CAP_* (cap masks)
2009-01-02 14:42:58 -08:00
Sage Weil
6ff545c099
mds: cap writeback authlock fields as well as filelock fields
...
No xattrs yet.
2009-01-02 14:26:08 -08:00
Sage Weil
7f46121db4
mds: share caps across multiple lock types.
...
Very rough initial first pass.
2009-01-02 12:05:09 -08:00
Sage Weil
4087a07478
todo
2009-01-02 09:26:27 -08:00
Sage Weil
7ad7e044ef
mds: clean up lock print methods
2009-01-01 11:06:13 -08:00
Sage Weil
a58d31ca46
Merge branch 'unstable' into asyncmds
...
Conflicts:
src/mon/OSDMonitor.cc
2008-12-31 11:30:08 -08:00
Sage Weil
142e2f6ef9
objecter: fix ack vs safe behavior
...
This was totally broken (ack was taken as a safe). Doh!
2008-12-31 11:27:07 -08:00
Sage Weil
0faa47e089
journaler: add --journaler_safe hack support back
2008-12-31 11:21:21 -08:00
Sage Weil
40277f652a
objecter: fix ack vs safe behavior
...
This was totally broken (ack was taken as a safe). Doh!
2008-12-31 11:20:29 -08:00
Sage Weil
9a3988e4f7
kclient: setattr on inode (not path) if holding WR cap
...
For newly created files, this avoids contention in the MDS over the dentry
xlock (e.g., for an untar).
It is slightly risky, because we could race with a caps release and MDS
cache trim. Pretty unlikely, however, because cap release would also
involve MDS journaling.
2008-12-31 11:19:38 -08:00
Sage Weil
4377dba319
mds: fix xattr projection
...
We do not want the new xattr to appear in the inode until the journal
entry commits.
2008-12-31 09:15:34 -08:00
Yehuda Sadeh
a77b0cbdc0
kclient: update for ceph_fsid_t
2008-12-30 16:18:01 -08:00
Yehuda Sadeh
af715a3a19
Merge branch 'unstable' of ssh://ceph.newdream.net/git/ceph into unstable
2008-12-24 16:39:59 -08:00
Yehuda Sadeh
d359afd3b7
vstart.sh: disable lockdep in default
2008-12-24 16:39:28 -08:00
Sage Weil
f46c77111d
mds todos
2008-12-24 14:40:23 -08:00
Sage Weil
3b253bdf9f
mds: avoid recursion on log trim
2008-12-24 14:19:31 -08:00
Sage Weil
99fc60e82c
mds: add wrlock to simplelock; wrlock authlock on chmod, chown
2008-12-24 14:19:30 -08:00
Sage Weil
3d93ca0143
mds: link on commit in openc; make find_snaprealm() follow projected_parent for unlinked inode
...
Linking inode in early confuses CDir::_committed().
2008-12-24 14:19:30 -08:00
Sage Weil
836513edcc
mon: make 'osd setmap' take epoch, to avoid executing multiple times
2008-12-24 14:19:27 -08:00
Sage Weil
6f9132e9bc
ceph: increase command timeout
2008-12-24 12:45:07 -08:00
Sage Weil
3ed25a3b31
mds: flush log only when we need to
...
In particular, don't flush immediately on cap flush unless the lock is
unstable. If we later wait on the lock state, make sure the log flushes.
2008-12-24 12:11:50 -08:00
Sage Weil
89900f30d6
mds: add --mds_early_reply option
2008-12-23 16:36:27 -08:00
Sage Weil
270a12a64e
mds: only include trace in first reply
...
No need to do it twice.
2008-12-23 16:36:12 -08:00
Sage Weil
838024464c
mds: rewrite, clean up mds open+O_CREAT; issue dn lease to xlocker; many other server cleanups
2008-12-23 15:43:10 -08:00
Sage Weil
8ba7f14c37
change ceph_fsid
...
Make it look more like uuid_t. Rename type, comparator.
2008-12-23 13:39:25 -08:00
Sage Weil
74cf3e7c42
mds: early_reply on open, mknod, symlink, mkdir when possible
...
Use client-provided ino, if defined.
2008-12-23 13:36:46 -08:00
Sage Weil
1419f00218
os: start_sync osd operation
...
If a client has a time sensitive operation, it can include a start_sync
in the op vector to start an immediate commit to disk. It should then
get the ONDISK ack sooner.
2008-12-23 12:29:04 -08:00
Sage Weil
3ed4b78556
osd: use .version in eversion_t to determine divergence
2008-12-23 12:19:11 -08:00
Sage Weil
b4d9cb59d6
osd: cleanup
2008-12-23 12:19:10 -08:00
