Sage Weil
c7da2e8768
msgr: fixed bug in accept
...
banner[] isn't null terminated
2008-10-01 15:34:16 -07:00
Sage Weil
73c689cb5d
messenger protocol changes
...
There is now a plaintext banner that is exchanged so that both ends can
verify they are talking to the right software. This string should be
changed each time the wire protocol changes.
There are also mon, osd, and mds specific protocol versions defined, but
these are not yet used.
2008-10-01 15:03:50 -07:00
Yehuda Sadeh
5ca0b3305d
kclient: whitespaces
2008-10-01 14:03:05 -07:00
Yehuda Sadeh
960bb0fb05
kclient: can supply multiple monitor ips on mount
2008-10-01 14:03:05 -07:00
Yehuda Sadeh
2b449c0807
kclient: adding unmaskable debug flag
2008-10-01 14:03:05 -07:00
Sage Weil
4893720941
makefile update
2008-10-01 12:00:53 -07:00
Sage Weil
2909ef79ce
v0.4, todos
2008-10-01 11:53:38 -07:00
Sage Weil
024d0c0766
config: comment
2008-10-01 11:40:48 -07:00
Sage Weil
1b3a1c4fe3
mds: fix file recovery
...
Make state bits distinct. Duh.
Also, if an inode is queued for recovery, and doesn't need it later, unqueue it
(clear the bit, drop auth pin).
2008-10-01 11:39:42 -07:00
Sage Weil
2542192741
kclient: fix pending cap_snap checks, queuing
...
Add new cap_snaps to the _tail_ of the list.
Also, use a helper to check for pending (write in progress) check, and be
sure to check the list tail in that case.
2008-10-01 10:50:51 -07:00
Sage Weil
51e0ff5823
client: fix queued cap_snap write-in-progress checks
2008-10-01 10:50:46 -07:00
Sage Weil
66daf9cdbc
kclient: be more paranoid in prepare_pages
...
It's theoretically possible to get racing incoming osd_op_replies. Make sure we only give the pages away once.
2008-09-30 12:59:27 -07:00
Sage Weil
1c886b78cc
kclient: fix bad pointer arithmetic in messenger
...
This was causing intermittent slab corruption when TCP fragmented the stream in unusual ways.
2008-09-30 12:58:17 -07:00
Sage Weil
91dd8ebed5
kclient: don't include pages in read requests; some osdc cleanup
2008-09-29 16:30:54 -07:00
Sage Weil
38770d2249
kclient: switch to sync writes when NEARFULL osdmap flag set; ENOSPC on FULL flag
2008-09-29 15:25:07 -07:00
Sage Weil
750c94411e
todos
2008-09-29 13:52:28 -07:00
Sage Weil
e3ce64fb71
kclient: default to safe writeback
...
Now that async writepages is working, we can wait for data to be fully committed without the higher latency affecting writeback throughput.
2008-09-29 13:31:24 -07:00
Sage Weil
aea485a2c1
mds: choose new mon only if beacons are delayed, and switch to mon that sends up beacons
...
This streamlines the mds <-> mon beacon interaction a bit.
2008-09-29 13:30:35 -07:00
Sage Weil
76c7cd0acb
kclient: some async writeback cleanup
2008-09-29 12:57:55 -07:00
Sage Weil
37d7bf5cd4
kclient: async writepages
2008-09-29 11:49:17 -07:00
Yehuda Sadeh
21650223d0
kclient: using special protocol mask for mds requests
2008-09-29 11:08:36 -07:00
Yehuda Sadeh
a4b26a3655
kclient: added missing file
2008-09-29 09:57:32 -07:00
Yehuda Sadeh
9f630cfead
kclient: mask to enable/disable debug messages
2008-09-29 09:54:48 -07:00
Sage Weil
8833e12e06
journal: fix off-by-one bug in committed_thru callbacks
2008-09-22 14:38:14 -07:00
Sage Weil
fd439bfaf2
journal: avoid queueing null context
2008-09-22 14:38:08 -07:00
Sage Weil
c623948359
kclient: writepage noop should be kswapd
2008-09-22 11:05:32 -07:00
Sage Weil
0d6f98f1d4
osdmaptool: print osdmap
2008-09-20 15:02:04 -07:00
Sage Weil
3d618f84f7
osdmap: set ctime+mtime on creation
2008-09-20 15:01:51 -07:00
Sage Weil
534746c76e
kclient: writeback cleanups (use filemap_write_and_wait)
2008-09-20 14:11:03 -07:00
Sage Weil
8ecd757767
kclient: time out mds session close on umount. messily.
...
This at least won't hang, but it will leave inodes pinned after umount, so it still needs some real work.
2008-09-20 13:51:12 -07:00
Sage Weil
c67f2c21ef
kclient: fix up safe/unsafe writeback flags
2008-09-20 13:37:11 -07:00
Sage Weil
17128988e1
kclient: unify mount timeouts
...
Make both mon map and mount messages and mds session, root inode open time out. Previously the mon interaction would time out but we could hang on a non-responsive (e.g., not active) mds.
2008-09-20 13:34:11 -07:00
Sage Weil
a0eb5e7d91
todos
2008-09-18 15:39:54 -07:00
Sage Weil
f8ec9d6c03
psim: test localized placement groups
2008-09-18 14:30:47 -07:00
Sage Weil
997a5f9e62
crush: fix forcefeeding for forceleaf
2008-09-18 14:30:01 -07:00
Sage Weil
3b5f237cd7
osdmaptool: lpg_bits specifiable
2008-09-18 14:26:57 -07:00
Sage Weil
ecf1ba1ad8
set bl length in buffer::read_file
2008-09-18 13:44:29 -07:00
Sage Weil
8adb5dd168
osdmaptool: fix crush import
2008-09-18 13:43:51 -07:00
Sage Weil
52f9d15826
crushtool: fix chooseleaf compilation/decompilation
2008-09-18 13:10:14 -07:00
Sage Weil
fb8cf71e27
todo
2008-09-17 15:05:35 -07:00
Sage Weil
0a9c9868e1
crush: chooseleaf
2008-09-17 14:39:05 -07:00
Sage Weil
8bcbcd6ba5
kclient: remove extraneous dput in fill_trace
2008-09-16 14:44:48 -07:00
Sage Weil
43e7e39972
osd: push/pull data_subset, clone_subsets
2008-09-16 14:44:47 -07:00
Sage Weil
1d20e73dad
objectstore: clonerange, list->vector encoding
2008-09-16 14:44:47 -07:00
Yehuda Sadeh
c43adbe5d6
kclient: lowered debug level on some message
2008-09-15 11:29:17 -07:00
Sage Weil
d76ce49140
kclient: fix mds reconnect message encoding; include snaprealms
2008-09-12 15:09:34 -07:00
Sage Weil
d7a103f596
filejournal: typo, fixes callbacks
2008-09-12 13:30:14 -07:00
Sage Weil
9ec27ab1c7
journal: remove some dead code
2008-09-12 12:51:10 -07:00
Sage Weil
07c3f0138f
kclient: fill in kstatfs f_fsid
2008-09-12 12:46:43 -07:00
Sage Weil
ae4b5ac30e
paxosservice: reset propose timer when election is called
2008-09-12 12:46:31 -07:00
