4 - fix removal of OSCs from LOV when they fail
5 - fix NULL deref during bulk timeout (214)
7 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
10 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
11 - defer open object destruction to close time (601981, 138)
12 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
13 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
14 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
15 - don't call obd_setattr in ll_file_release for destroyed objects
16 * protocol change to lustre_msg: move |version| and add |flags|
17 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
18 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
19 * added replay of create, unlink, link and rename operations during
20 MDS failover; recovery should be much more robust now
21 * remove failed OSCs from LOVs (only lov_create uses this so far)
22 * the lustre-HOWTO was brought (more) up to date (582544)
24 2002-10-23 Phil Schwan <phil@clusterfs.com>
27 - in-use dentries weren't being reused properly (617851)
28 - prevent multiple LDLM setup (599178)
29 - fix LOV size calculations for truncate (617853)
30 - fix client handling of MDS intent errors (POSIX)
31 - fix permission bug in lovstripe.c test (624321)
32 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
33 - truncate past end of file could corrupt data
34 - proper cleanup after timeouts, crashes, etc (592524, 550815)
35 - a race in recovery could return ETIMEDOUT to apps (623947)
36 - building outside the source directory was fixed
37 * the lustre-HOWTO was brought (more) up to date (582544)
38 * major progress was made on recovery functionality
40 2002-10-10 Phil Schwan <phil@clusterfs.com>
43 - recovery deadlock fix
44 - rm -rf causes LBUG fix (617817)
45 - file open by multiple tasks fix (618962)
46 - directory permissions bugs (602707 and 620007)
47 - journal_stop fixed with locking (611313)
48 - O_APPEND failures resolved (618273, perhaps 614459)
49 - lconf PATH fix (619770)
50 - IA64 build fix (621450)
51 - RPC buffer sizes scale with amount of memory
53 2002-10-01 Phil Schwan <phil@clusterfs.com>
56 - locks would be cancelled without throwing away data pages,
57 resulting in inconsistent data (605627)
58 - inode attributes were not always being refreshed (605627, 612449)
59 - lconf now continues to cleanup after lctl reports an error
60 - MDS now enforces user permissions (602707)
61 - lprocfs cleanup fixed, but not yet enabled (614157)
62 - fixed infinite server hang, should a client not respond to an AST
63 - avoid going into recovery if user calls readlink() with a buffer
64 that's too small (613941)
65 - AST RPCs no longer require replies (614867) -- this may be changed
66 - don't crash server if client sends an IOV that's too big (611336)
67 - fixed lock conversion deadlock (611892)
68 - fixed the following of symlinks (614622)
69 * recovery: the server can remove locks from a client that dies, other
70 clients can make progress
71 * more extN patch fixes
72 * compile-time configurable ptlrpc buffer allocations
74 - collaborative read cache document
75 - Lustre Lite Performance CDR document-in-progress
77 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
80 - fix typo in patch-2.4.18
82 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
85 - clear ptlrpc request each time in handle_incoming_request()
86 - unlink of files now destroys the object on the OST
88 2002-09-19 Peter Braam <braam@clusterfs.com>
90 * add hard link support
91 * change obdfile creation method
92 * kernel patch changed
94 2002-09-19 Peter Braam <braam@clusterfs.com>
97 - stack overflow bug in extN fixed
99 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
101 * documentation updates
102 - add man pages for config tools
103 - update tests/README to describe testing with new config tools
104 - finish metadata API descriptions
105 * bug fixes and cleanups
106 - statfs workaround for 16TB limit
107 - LOV stripe allocation improved, can stripe on subset of OSTs
108 - LOV file size/IO offset was wrong for files > 4GB in size
109 - object EA data was being dropped, caused files to be unreadable
110 - memory overflow with non-LOV OST caused memory corruption
111 - fixed regression tests to work with new config tools, obdfilter
112 - fixed bug when directory size became larger than 1 block
113 - fixed bug (for single client case) when PWD was deleted
114 - invalidate local directory pages when doing intent-based ops
115 - avoid LDLM oops when lock callback contained bad data
117 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
119 * documentation updates
120 * bug fixes and cleanups
121 - configuration tools
124 - 64-bit compile warnings
125 - 64-bit internal statfs data
127 * test_brw on persistent OST devices
131 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
133 * documentation updates
134 * bug fixes and cleanups
135 * configuration tools
137 2002-08-30 Peter J. Braam <braam@clusterfs.com>
140 * many small fixes to 0_5_4
141 * io/network handling
142 * thinkos in MDS operations
144 2002-08-24 Peter J. Braam <braam@clusterfs.com>
147 * crucial basic fixes to 0.5.3
148 * IOR, Iozone work over Elan
151 2002-08-07 Phil Schwan <phil@clusterfs.com>
152 * version 0_5_3, our first alpha
153 * we use the new Portals iovs
154 * documentation updates
155 * bug fixes and cleanups
156 * small changes in the DLM wire protocol
158 2002-07-25 Peter J. Braam <braam@clusterfs.com>
159 * version 0_5_1 with some initial stability,
160 * locking on MD and file I/O.
161 * documentation updates
162 * several bug fixes since 0.5.0
163 * small changes in wire protocol
165 2002-07-18 Phil Schwan <phil@clusterfs.com>
167 * delivered as Lustre Light Alpha
168 * fixed a crash after handling invalid MDS requests
169 * fixed directory pages for architectures with non-4k pages sizes
171 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
172 * release version v0_4_4
173 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
175 * Fixes a number of interruption problems with OST operations.
176 * Update documentation for portals header changes
177 * Move all wire protocol structs/defines to lustre_idl.h
178 * Fixes symlink length bug.
179 * Add tcpdump to repository.
181 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
182 * release version v0_4_3
183 * Fixes statfs for inodes on extN.
184 * Fixes bug in runtests which would delete /etc/hosts.
185 * Use 64-bit object IDs wherever possible (not into VFS though)
186 Remove ost_get_info, which is unused by lustre, and out of date.
188 2002-07-03 Peter Braam <braam@clusterfs.com>
189 * release version v0_4_2 Fixes a lookup error (type not passed)
190 * move forward to head of Portals
191 * move forward to latest Lustre kernel
193 2002-06-25 Peter Braam <braam@clusterfs.com>
194 * release version v0_4_1. Hopefully stable on single node use.