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)
17 - added a maximum lock count, an LRU list, and a flusher
19 - Add capability to MDS protocol
20 - LDLM cancellations and callbacks on different portals
22 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
25 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
26 - defer open object destruction to close time (601981, 138)
27 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
28 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
29 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
30 - don't call obd_setattr in ll_file_release for destroyed objects
31 * protocol change to lustre_msg: move |version| and add |flags|
32 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
33 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
34 * added replay of create, unlink, link and rename operations during
35 MDS failover; recovery should be much more robust now
36 * remove failed OSCs from LOVs (only lov_create uses this so far)
37 * the lustre-HOWTO was brought (more) up to date (582544)
39 2002-10-23 Phil Schwan <phil@clusterfs.com>
42 - in-use dentries weren't being reused properly (617851)
43 - prevent multiple LDLM setup (599178)
44 - fix LOV size calculations for truncate (617853)
45 - fix client handling of MDS intent errors (POSIX)
46 - fix permission bug in lovstripe.c test (624321)
47 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
48 - truncate past end of file could corrupt data
49 - proper cleanup after timeouts, crashes, etc (592524, 550815)
50 - a race in recovery could return ETIMEDOUT to apps (623947)
51 - building outside the source directory was fixed
52 * the lustre-HOWTO was brought (more) up to date (582544)
53 * major progress was made on recovery functionality
55 2002-10-10 Phil Schwan <phil@clusterfs.com>
58 - recovery deadlock fix
59 - rm -rf causes LBUG fix (617817)
60 - file open by multiple tasks fix (618962)
61 - directory permissions bugs (602707 and 620007)
62 - journal_stop fixed with locking (611313)
63 - O_APPEND failures resolved (618273, perhaps 614459)
64 - lconf PATH fix (619770)
65 - IA64 build fix (621450)
66 - RPC buffer sizes scale with amount of memory
68 2002-10-01 Phil Schwan <phil@clusterfs.com>
71 - locks would be cancelled without throwing away data pages,
72 resulting in inconsistent data (605627)
73 - inode attributes were not always being refreshed (605627, 612449)
74 - lconf now continues to cleanup after lctl reports an error
75 - MDS now enforces user permissions (602707)
76 - lprocfs cleanup fixed, but not yet enabled (614157)
77 - fixed infinite server hang, should a client not respond to an AST
78 - avoid going into recovery if user calls readlink() with a buffer
79 that's too small (613941)
80 - AST RPCs no longer require replies (614867) -- this may be changed
81 - don't crash server if client sends an IOV that's too big (611336)
82 - fixed lock conversion deadlock (611892)
83 - fixed the following of symlinks (614622)
84 * recovery: the server can remove locks from a client that dies, other
85 clients can make progress
86 * more extN patch fixes
87 * compile-time configurable ptlrpc buffer allocations
89 - collaborative read cache document
90 - Lustre Lite Performance CDR document-in-progress
92 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
95 - fix typo in patch-2.4.18
97 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
100 - clear ptlrpc request each time in handle_incoming_request()
101 - unlink of files now destroys the object on the OST
103 2002-09-19 Peter Braam <braam@clusterfs.com>
105 * add hard link support
106 * change obdfile creation method
107 * kernel patch changed
109 2002-09-19 Peter Braam <braam@clusterfs.com>
112 - stack overflow bug in extN fixed
114 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
116 * documentation updates
117 - add man pages for config tools
118 - update tests/README to describe testing with new config tools
119 - finish metadata API descriptions
120 * bug fixes and cleanups
121 - statfs workaround for 16TB limit
122 - LOV stripe allocation improved, can stripe on subset of OSTs
123 - LOV file size/IO offset was wrong for files > 4GB in size
124 - object EA data was being dropped, caused files to be unreadable
125 - memory overflow with non-LOV OST caused memory corruption
126 - fixed regression tests to work with new config tools, obdfilter
127 - fixed bug when directory size became larger than 1 block
128 - fixed bug (for single client case) when PWD was deleted
129 - invalidate local directory pages when doing intent-based ops
130 - avoid LDLM oops when lock callback contained bad data
132 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
134 * documentation updates
135 * bug fixes and cleanups
136 - configuration tools
139 - 64-bit compile warnings
140 - 64-bit internal statfs data
142 * test_brw on persistent OST devices
146 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
148 * documentation updates
149 * bug fixes and cleanups
150 * configuration tools
152 2002-08-30 Peter J. Braam <braam@clusterfs.com>
155 * many small fixes to 0_5_4
156 * io/network handling
157 * thinkos in MDS operations
159 2002-08-24 Peter J. Braam <braam@clusterfs.com>
162 * crucial basic fixes to 0.5.3
163 * IOR, Iozone work over Elan
166 2002-08-07 Phil Schwan <phil@clusterfs.com>
167 * version 0_5_3, our first alpha
168 * we use the new Portals iovs
169 * documentation updates
170 * bug fixes and cleanups
171 * small changes in the DLM wire protocol
173 2002-07-25 Peter J. Braam <braam@clusterfs.com>
174 * version 0_5_1 with some initial stability,
175 * locking on MD and file I/O.
176 * documentation updates
177 * several bug fixes since 0.5.0
178 * small changes in wire protocol
180 2002-07-18 Phil Schwan <phil@clusterfs.com>
182 * delivered as Lustre Light Alpha
183 * fixed a crash after handling invalid MDS requests
184 * fixed directory pages for architectures with non-4k pages sizes
186 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
187 * release version v0_4_4
188 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
190 * Fixes a number of interruption problems with OST operations.
191 * Update documentation for portals header changes
192 * Move all wire protocol structs/defines to lustre_idl.h
193 * Fixes symlink length bug.
194 * Add tcpdump to repository.
196 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
197 * release version v0_4_3
198 * Fixes statfs for inodes on extN.
199 * Fixes bug in runtests which would delete /etc/hosts.
200 * Use 64-bit object IDs wherever possible (not into VFS though)
201 Remove ost_get_info, which is unused by lustre, and out of date.
203 2002-07-03 Peter Braam <braam@clusterfs.com>
204 * release version v0_4_2 Fixes a lookup error (type not passed)
205 * move forward to head of Portals
206 * move forward to latest Lustre kernel
208 2002-06-25 Peter Braam <braam@clusterfs.com>
209 * release version v0_4_1. Hopefully stable on single node use.