4 - fix null d_it dereference (346)
5 - fix full OST/dbench hang (333)
6 - fix permission problem with file removal (286)
7 - fix removal of OSCs from LOV when they fail
8 - fix NULL deref during bulk timeout (214)
9 - fix problems related to multiple filesystems on one MDS (241)
10 - fixed serious subtle metadata locking bugs
11 - free locks on clients when inodes are removed due to memory
13 - fix inode pointer in lock data (285)
14 - partial support for multiple MDS on a single host (623279, 241)
15 - data locks weren't cancelled at clear_inode time (290, 311)
16 - intent locks could lead to unbounded lock growth (205)
18 - Add capability to MDS protocol
19 - LDLM cancellations and callbacks on different portals
21 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
24 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
25 - defer open object destruction to close time (601981, 138)
26 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
27 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
28 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
29 - don't call obd_setattr in ll_file_release for destroyed objects
30 * protocol change to lustre_msg: move |version| and add |flags|
31 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
32 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
33 * added replay of create, unlink, link and rename operations during
34 MDS failover; recovery should be much more robust now
35 * remove failed OSCs from LOVs (only lov_create uses this so far)
36 * the lustre-HOWTO was brought (more) up to date (582544)
38 2002-10-23 Phil Schwan <phil@clusterfs.com>
41 - in-use dentries weren't being reused properly (617851)
42 - prevent multiple LDLM setup (599178)
43 - fix LOV size calculations for truncate (617853)
44 - fix client handling of MDS intent errors (POSIX)
45 - fix permission bug in lovstripe.c test (624321)
46 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
47 - truncate past end of file could corrupt data
48 - proper cleanup after timeouts, crashes, etc (592524, 550815)
49 - a race in recovery could return ETIMEDOUT to apps (623947)
50 - building outside the source directory was fixed
51 * the lustre-HOWTO was brought (more) up to date (582544)
52 * major progress was made on recovery functionality
54 2002-10-10 Phil Schwan <phil@clusterfs.com>
57 - recovery deadlock fix
58 - rm -rf causes LBUG fix (617817)
59 - file open by multiple tasks fix (618962)
60 - directory permissions bugs (602707 and 620007)
61 - journal_stop fixed with locking (611313)
62 - O_APPEND failures resolved (618273, perhaps 614459)
63 - lconf PATH fix (619770)
64 - IA64 build fix (621450)
65 - RPC buffer sizes scale with amount of memory
67 2002-10-01 Phil Schwan <phil@clusterfs.com>
70 - locks would be cancelled without throwing away data pages,
71 resulting in inconsistent data (605627)
72 - inode attributes were not always being refreshed (605627, 612449)
73 - lconf now continues to cleanup after lctl reports an error
74 - MDS now enforces user permissions (602707)
75 - lprocfs cleanup fixed, but not yet enabled (614157)
76 - fixed infinite server hang, should a client not respond to an AST
77 - avoid going into recovery if user calls readlink() with a buffer
78 that's too small (613941)
79 - AST RPCs no longer require replies (614867) -- this may be changed
80 - don't crash server if client sends an IOV that's too big (611336)
81 - fixed lock conversion deadlock (611892)
82 - fixed the following of symlinks (614622)
83 * recovery: the server can remove locks from a client that dies, other
84 clients can make progress
85 * more extN patch fixes
86 * compile-time configurable ptlrpc buffer allocations
88 - collaborative read cache document
89 - Lustre Lite Performance CDR document-in-progress
91 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
94 - fix typo in patch-2.4.18
96 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
99 - clear ptlrpc request each time in handle_incoming_request()
100 - unlink of files now destroys the object on the OST
102 2002-09-19 Peter Braam <braam@clusterfs.com>
104 * add hard link support
105 * change obdfile creation method
106 * kernel patch changed
108 2002-09-19 Peter Braam <braam@clusterfs.com>
111 - stack overflow bug in extN fixed
113 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
115 * documentation updates
116 - add man pages for config tools
117 - update tests/README to describe testing with new config tools
118 - finish metadata API descriptions
119 * bug fixes and cleanups
120 - statfs workaround for 16TB limit
121 - LOV stripe allocation improved, can stripe on subset of OSTs
122 - LOV file size/IO offset was wrong for files > 4GB in size
123 - object EA data was being dropped, caused files to be unreadable
124 - memory overflow with non-LOV OST caused memory corruption
125 - fixed regression tests to work with new config tools, obdfilter
126 - fixed bug when directory size became larger than 1 block
127 - fixed bug (for single client case) when PWD was deleted
128 - invalidate local directory pages when doing intent-based ops
129 - avoid LDLM oops when lock callback contained bad data
131 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
133 * documentation updates
134 * bug fixes and cleanups
135 - configuration tools
138 - 64-bit compile warnings
139 - 64-bit internal statfs data
141 * test_brw on persistent OST devices
145 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
147 * documentation updates
148 * bug fixes and cleanups
149 * configuration tools
151 2002-08-30 Peter J. Braam <braam@clusterfs.com>
154 * many small fixes to 0_5_4
155 * io/network handling
156 * thinkos in MDS operations
158 2002-08-24 Peter J. Braam <braam@clusterfs.com>
161 * crucial basic fixes to 0.5.3
162 * IOR, Iozone work over Elan
165 2002-08-07 Phil Schwan <phil@clusterfs.com>
166 * version 0_5_3, our first alpha
167 * we use the new Portals iovs
168 * documentation updates
169 * bug fixes and cleanups
170 * small changes in the DLM wire protocol
172 2002-07-25 Peter J. Braam <braam@clusterfs.com>
173 * version 0_5_1 with some initial stability,
174 * locking on MD and file I/O.
175 * documentation updates
176 * several bug fixes since 0.5.0
177 * small changes in wire protocol
179 2002-07-18 Phil Schwan <phil@clusterfs.com>
181 * delivered as Lustre Light Alpha
182 * fixed a crash after handling invalid MDS requests
183 * fixed directory pages for architectures with non-4k pages sizes
185 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
186 * release version v0_4_4
187 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
189 * Fixes a number of interruption problems with OST operations.
190 * Update documentation for portals header changes
191 * Move all wire protocol structs/defines to lustre_idl.h
192 * Fixes symlink length bug.
193 * Add tcpdump to repository.
195 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
196 * release version v0_4_3
197 * Fixes statfs for inodes on extN.
198 * Fixes bug in runtests which would delete /etc/hosts.
199 * Use 64-bit object IDs wherever possible (not into VFS though)
200 Remove ost_get_info, which is unused by lustre, and out of date.
202 2002-07-03 Peter Braam <braam@clusterfs.com>
203 * release version v0_4_2 Fixes a lookup error (type not passed)
204 * move forward to head of Portals
205 * move forward to latest Lustre kernel
207 2002-06-25 Peter Braam <braam@clusterfs.com>
208 * release version v0_4_1. Hopefully stable on single node use.