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
18 - fix multiple rename (365)
20 - Add capability to MDS protocol
21 - LDLM cancellations and callbacks on different portals
23 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
26 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
27 - defer open object destruction to close time (601981, 138)
28 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
29 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
30 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
31 - don't call obd_setattr in ll_file_release for destroyed objects
32 * protocol change to lustre_msg: move |version| and add |flags|
33 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
34 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
35 * added replay of create, unlink, link and rename operations during
36 MDS failover; recovery should be much more robust now
37 * remove failed OSCs from LOVs (only lov_create uses this so far)
38 * the lustre-HOWTO was brought (more) up to date (582544)
40 2002-10-23 Phil Schwan <phil@clusterfs.com>
43 - in-use dentries weren't being reused properly (617851)
44 - prevent multiple LDLM setup (599178)
45 - fix LOV size calculations for truncate (617853)
46 - fix client handling of MDS intent errors (POSIX)
47 - fix permission bug in lovstripe.c test (624321)
48 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
49 - truncate past end of file could corrupt data
50 - proper cleanup after timeouts, crashes, etc (592524, 550815)
51 - a race in recovery could return ETIMEDOUT to apps (623947)
52 - building outside the source directory was fixed
53 * the lustre-HOWTO was brought (more) up to date (582544)
54 * major progress was made on recovery functionality
56 2002-10-10 Phil Schwan <phil@clusterfs.com>
59 - recovery deadlock fix
60 - rm -rf causes LBUG fix (617817)
61 - file open by multiple tasks fix (618962)
62 - directory permissions bugs (602707 and 620007)
63 - journal_stop fixed with locking (611313)
64 - O_APPEND failures resolved (618273, perhaps 614459)
65 - lconf PATH fix (619770)
66 - IA64 build fix (621450)
67 - RPC buffer sizes scale with amount of memory
69 2002-10-01 Phil Schwan <phil@clusterfs.com>
72 - locks would be cancelled without throwing away data pages,
73 resulting in inconsistent data (605627)
74 - inode attributes were not always being refreshed (605627, 612449)
75 - lconf now continues to cleanup after lctl reports an error
76 - MDS now enforces user permissions (602707)
77 - lprocfs cleanup fixed, but not yet enabled (614157)
78 - fixed infinite server hang, should a client not respond to an AST
79 - avoid going into recovery if user calls readlink() with a buffer
80 that's too small (613941)
81 - AST RPCs no longer require replies (614867) -- this may be changed
82 - don't crash server if client sends an IOV that's too big (611336)
83 - fixed lock conversion deadlock (611892)
84 - fixed the following of symlinks (614622)
85 * recovery: the server can remove locks from a client that dies, other
86 clients can make progress
87 * more extN patch fixes
88 * compile-time configurable ptlrpc buffer allocations
90 - collaborative read cache document
91 - Lustre Lite Performance CDR document-in-progress
93 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
96 - fix typo in patch-2.4.18
98 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
101 - clear ptlrpc request each time in handle_incoming_request()
102 - unlink of files now destroys the object on the OST
104 2002-09-19 Peter Braam <braam@clusterfs.com>
106 * add hard link support
107 * change obdfile creation method
108 * kernel patch changed
110 2002-09-19 Peter Braam <braam@clusterfs.com>
113 - stack overflow bug in extN fixed
115 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
117 * documentation updates
118 - add man pages for config tools
119 - update tests/README to describe testing with new config tools
120 - finish metadata API descriptions
121 * bug fixes and cleanups
122 - statfs workaround for 16TB limit
123 - LOV stripe allocation improved, can stripe on subset of OSTs
124 - LOV file size/IO offset was wrong for files > 4GB in size
125 - object EA data was being dropped, caused files to be unreadable
126 - memory overflow with non-LOV OST caused memory corruption
127 - fixed regression tests to work with new config tools, obdfilter
128 - fixed bug when directory size became larger than 1 block
129 - fixed bug (for single client case) when PWD was deleted
130 - invalidate local directory pages when doing intent-based ops
131 - avoid LDLM oops when lock callback contained bad data
133 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
135 * documentation updates
136 * bug fixes and cleanups
137 - configuration tools
140 - 64-bit compile warnings
141 - 64-bit internal statfs data
143 * test_brw on persistent OST devices
147 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
149 * documentation updates
150 * bug fixes and cleanups
151 * configuration tools
153 2002-08-30 Peter J. Braam <braam@clusterfs.com>
156 * many small fixes to 0_5_4
157 * io/network handling
158 * thinkos in MDS operations
160 2002-08-24 Peter J. Braam <braam@clusterfs.com>
163 * crucial basic fixes to 0.5.3
164 * IOR, Iozone work over Elan
167 2002-08-07 Phil Schwan <phil@clusterfs.com>
168 * version 0_5_3, our first alpha
169 * we use the new Portals iovs
170 * documentation updates
171 * bug fixes and cleanups
172 * small changes in the DLM wire protocol
174 2002-07-25 Peter J. Braam <braam@clusterfs.com>
175 * version 0_5_1 with some initial stability,
176 * locking on MD and file I/O.
177 * documentation updates
178 * several bug fixes since 0.5.0
179 * small changes in wire protocol
181 2002-07-18 Phil Schwan <phil@clusterfs.com>
183 * delivered as Lustre Light Alpha
184 * fixed a crash after handling invalid MDS requests
185 * fixed directory pages for architectures with non-4k pages sizes
187 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
188 * release version v0_4_4
189 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
191 * Fixes a number of interruption problems with OST operations.
192 * Update documentation for portals header changes
193 * Move all wire protocol structs/defines to lustre_idl.h
194 * Fixes symlink length bug.
195 * Add tcpdump to repository.
197 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
198 * release version v0_4_3
199 * Fixes statfs for inodes on extN.
200 * Fixes bug in runtests which would delete /etc/hosts.
201 * Use 64-bit object IDs wherever possible (not into VFS though)
202 Remove ost_get_info, which is unused by lustre, and out of date.
204 2002-07-03 Peter Braam <braam@clusterfs.com>
205 * release version v0_4_2 Fixes a lookup error (type not passed)
206 * move forward to head of Portals
207 * move forward to latest Lustre kernel
209 2002-06-25 Peter Braam <braam@clusterfs.com>
210 * release version v0_4_1. Hopefully stable on single node use.