RepoMirrors/ceph
1
0
Fork 0
mirror of https://github.com/ceph/ceph synced 2024-12-14 07:25:50 +00:00
Code Issues Packages Projects Releases Wiki Activity
7,357 Commits 211 Branches 394 Tags 1.9 GiB
e9f456843c
Commit Graph

7357 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Sage Weil
e9f456843c msgr: fix race with signals vs select with pselect 
Fixes hang where accepter thread misses the stop signal and
doesn't shut down.
 2009-07-14 10:35:03 -07:00
Sage Weil
82fe78cf1d mkcephfs: chown/chmod mount point if non-root 2009-07-14 10:17:29 -07:00
Sage Weil
e4f817030f mds: reconnect caps added via replay+issue_new_caps 
Also, set up the initial lock states after reconnecting them!
 2009-07-13 17:48:19 -07:00
Sage Weil
ea2f8343d2 mds: check client ranges on mds restart 2009-07-13 17:48:19 -07:00
Sage Weil
2b03d81b1d mds: adjust client range on open existing for write 2009-07-13 17:48:19 -07:00
Sage Weil
e842cf55b2 mds: journal recovered mtime; cleanup 
Also avoid using in->inode as that may fail with concurrent updates
(if that's even possible?).
 2009-07-13 17:48:19 -07:00
Sage Weil
6867c0cb9a mds: do not check_inode_max_size in eval() path 
This is super dangerous, since we may journal something in the
middle of preparing another journal entry.  (This is notably a
problem with calls to issue_new_caps).

Kill fugly NO_SIZE_CHECK flag too, yay.
 2009-07-13 17:48:18 -07:00
Sage Weil
b1d5bea560 mds: set client_range on open O_TRUNC 
Set client_ranges explicitly.  Move get_layout_size_inc into
inode_t.
 2009-07-13 17:48:18 -07:00
Sage Weil
0b454296a3 mds: base client_ranges update on issued|wanted, not just issued 2009-07-13 17:48:18 -07:00
Sage Weil
23aeea4ee7 mds: use SCAN state for file recovery on mds restart 
This ensures we revoke other caps before scanning for file
sizes.
 2009-07-13 17:48:18 -07:00
Sage Weil
0d5b2c08ad Makefile: fix whitespace 2009-07-13 10:39:25 -07:00
Sage Weil
5cfa35e4af initscript: fix 'reload' 2009-07-13 10:38:23 -07:00
Sage Weil
9a7fdb22a9 initscripts: do what we can to run as user specified in the conf 2009-07-13 10:38:23 -07:00
Sage Weil
5bb2968985 initscripts: substitute in proper autoconf paths 
Based on
http://www.gnu.org/software/hello/manual/autoconf/Installation-Directory-Variables.html
 2009-07-13 10:30:43 -07:00
Sage Weil
08a6fb733d mds todos 2009-07-10 15:32:56 -07:00
Sage Weil
2789732c9e kclient: simplify/fix cap delay timeouts 
Do NOT reset timeouts at the top of ceph_check_caps.. that's just
wrong, and screws up cap release logic.

Reset timeouts when we requeue the cap, since that's what all the
callers do.
 2009-07-10 15:32:56 -07:00
Sage Weil
f871d3e0a0 mds: on restart, only recover files affected by missing clients 2009-07-10 15:32:56 -07:00
Sage Weil
3a86b94cb3 mds: maintain per-client write ranges instead of single max_size 
We need to track individual clients' ability to write to a file in
order to behave rationally when a client dies or fails to reconnect
after an mds restart.

There is also a behavior change hiding in here: before we were
adjusting max_size based on _wanted_ caps, but that was unsafe,
since a client may have the old max_size value and initiate a write
after informing the mds it no longer wants said caps.  There will
undoubtably be some fallout from that change...

Also clean up instances of memset() on inode_t and friends.  Define
and use constructors instead.
 2009-07-10 15:32:56 -07:00
Sage Weil
167f4cad26 uclient: fix getdir() to use new readdir_r return value 2009-07-10 15:32:55 -07:00
Sage Weil
8733b56fe5 uclient: stop using mds inode_t 2009-07-10 15:32:55 -07:00
Sage Weil
28e444dd0a mds: add SCAN, PRE_SCAN lock states 
Also clean up state machine array sizes
 2009-07-10 15:32:55 -07:00
Sage Weil
2ca5659eec osd: less hexdump on tmap updates 2009-07-10 15:32:55 -07:00
Yehuda Sadeh
bdfc852106 objecter: fix pool objects listing 2009-07-10 13:19:13 -07:00
Sage Weil
58fa7c1bed kclient: fix uninitialized vars 2009-07-09 16:44:09 -07:00
Sage Weil
21804bcc19 todo 2009-07-09 16:34:54 -07:00
Sage Weil
9a5d053459 Revert "kclient: strip out kernel version compatibility cruft" 
This reverts commit ac1cf5dc7343953462aa27dc3526dd5ccec9cb53.
 2009-07-09 16:33:38 -07:00
Sage Weil
28aa4e8a43 Merge branch 'master' of ssh://ceph.newdream.net/git/ceph 2009-07-09 16:29:36 -07:00
Sage Weil
67822ef05a s3gw: make compilation optional (./configure --with-s3gw) 2009-07-09 16:28:18 -07:00
Sage Weil
280e4df95d uclient: include tid in cap flush, flushsnap 2009-07-09 15:50:29 -07:00
Sage Weil
1961fb75eb kclient: include tid in cap flush, flushsnap 
This avoids (dirty 1, flush 1, dirty 2, flush 2, flush_ack 1 ->
mark clean) badness.
 2009-07-09 15:41:23 -07:00
Sage Weil
6d95e290aa mds: add tid to cap flush/flushsnap ops 
And include in the ACK.
 2009-07-09 15:35:22 -07:00
Sage Weil
45e409451a kclient: assign seq to cap flush; avoid starvation on sync_fs 
Assign a seq# to each cap flush.  When we sync_fs, wait only for
prior cap flushes to complete.

Also, fix wait_unsafe_requests.
 2009-07-09 15:11:52 -07:00
Sage Weil
c8744d6d48 kclient: debug output cleanup 2009-07-09 14:35:19 -07:00
Sage Weil
b5d32fa451 uclient: assign sequence to cap flushes, avoiding sync_fs starvation 
Order flushes, so we can wait for them without possibility of
starvation.
 2009-07-09 14:35:19 -07:00
Sage Weil
1a255ca1a0 kclient: fix write_inode; force dirty cap flush by mdsc_sync 
Adding flag to ceph_check_flags to flush any dirty data.  Pass
through from mdsc_sync.

Also, flush if i_ceph_flags & CEPH_I_FLUSH to avoid letting
write_inode calls similarly leak through.
 2009-07-09 14:35:19 -07:00
Greg Farnum
0bd233424b Merged Makefile.am 2009-07-09 14:06:43 -07:00
Greg Farnum
dab180c1ba libceph: SimpleMessenger usage bugfix, added testceph.cc, fixed dependencies 2009-07-09 14:04:06 -07:00
Sage Weil
a4b0764c94 kclient: make sync_fs wait on unsafe mds requests 2009-07-09 12:00:17 -07:00
Sage Weil
72b03df493 libceph: add dependencies to the Makefile 2009-07-09 11:54:22 -07:00
Greg Farnum
6c1fdddf34 libceph: Added getdir to the C++ interface. 
Used by hypertable
 2009-07-09 11:40:39 -07:00
Greg Farnum
4260265415 Hypertable:Completed initial Cephbroker. Requires testing. 2009-07-09 11:40:39 -07:00
Greg Farnum
523d7033ae hypertable: added skeleton .h and .cc CephBroker 
These are meant to be built in Hypertable, obviously.
 2009-07-09 11:40:38 -07:00
Sage Weil
0f6e0b3953 uclient: resend cap flushes on mds restart 
Fuse doesn't implement sync_fs, so this is difficult to test. :/
 2009-07-09 11:17:39 -07:00
Sage Weil
387778e6a0 uclient: change readdir_r return value 
1 on success (got a dirent), 0 on end of dir, <0 on error.  This
isn't exactly posix, but is a little less silly.
 2009-07-09 10:29:05 -07:00
Sage Weil
8ecffa5686 mds: make clientreplay go faster 
Queue up next op as soon as we submit the journal entry.  That maintains
ordering while avoiding any waiting for the journal to flush.
 2009-07-09 09:11:57 -07:00
Sage Weil
4daf4762cc uclient: track issue_seq; reset on reconnect 2009-07-08 16:31:58 -07:00
Sage Weil
2acd03a6fb kclient: reset issue_seq on reconnect 2009-07-08 16:31:39 -07:00
Sage Weil
1d4deb1296 mds: ignore cap updates in replayed ops 
We already have the caps they think they have; these releases are
obsolete.
 2009-07-08 16:31:25 -07:00
Sage Weil
cc54b17150 mds: do not include trace with replayed replies 2009-07-08 16:30:41 -07:00
Sage Weil
5b2b455e24 kclient: track and kick flushing caps per-mds 
After an MDS restarts, we need to resend cap flush messages we
were trying to write back.  So, track flushing caps per-session,
with a global counter to make ceph_mdsc_sync simple.  When the
MDS goes active, re-flush anything that was flushing.

Also change old 'sync' terminology to more descriptive 'flushing'.
 2009-07-08 15:55:03 -07:00