4 - fix removal of OSCs from LOV when they fail
5 - fix NULL deref during bulk timeout (214)
6 - fix problems related to multiple filesystems on one MDS (241)
7 - fixed serious subtle metadata locking bugs
8 - free locks on clients when inodes are removed due to memory
10 - fix inode pointer in lock data (285)
11 - partial support for multiple MDS on a single host (623279, 241)
13 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
16 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
17 - defer open object destruction to close time (601981, 138)
18 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
19 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
20 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
21 - don't call obd_setattr in ll_file_release for destroyed objects
22 * protocol change to lustre_msg: move |version| and add |flags|
23 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
24 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
25 * added replay of create, unlink, link and rename operations during
26 MDS failover; recovery should be much more robust now
27 * remove failed OSCs from LOVs (only lov_create uses this so far)
28 * the lustre-HOWTO was brought (more) up to date (582544)
30 2002-10-23 Phil Schwan <phil@clusterfs.com>
33 - in-use dentries weren't being reused properly (617851)
34 - prevent multiple LDLM setup (599178)
35 - fix LOV size calculations for truncate (617853)
36 - fix client handling of MDS intent errors (POSIX)
37 - fix permission bug in lovstripe.c test (624321)
38 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
39 - truncate past end of file could corrupt data
40 - proper cleanup after timeouts, crashes, etc (592524, 550815)
41 - a race in recovery could return ETIMEDOUT to apps (623947)
42 - building outside the source directory was fixed
43 * the lustre-HOWTO was brought (more) up to date (582544)
44 * major progress was made on recovery functionality
46 2002-10-10 Phil Schwan <phil@clusterfs.com>
49 - recovery deadlock fix
50 - rm -rf causes LBUG fix (617817)
51 - file open by multiple tasks fix (618962)
52 - directory permissions bugs (602707 and 620007)
53 - journal_stop fixed with locking (611313)
54 - O_APPEND failures resolved (618273, perhaps 614459)
55 - lconf PATH fix (619770)
56 - IA64 build fix (621450)
57 - RPC buffer sizes scale with amount of memory
59 2002-10-01 Phil Schwan <phil@clusterfs.com>
62 - locks would be cancelled without throwing away data pages,
63 resulting in inconsistent data (605627)
64 - inode attributes were not always being refreshed (605627, 612449)
65 - lconf now continues to cleanup after lctl reports an error
66 - MDS now enforces user permissions (602707)
67 - lprocfs cleanup fixed, but not yet enabled (614157)
68 - fixed infinite server hang, should a client not respond to an AST
69 - avoid going into recovery if user calls readlink() with a buffer
70 that's too small (613941)
71 - AST RPCs no longer require replies (614867) -- this may be changed
72 - don't crash server if client sends an IOV that's too big (611336)
73 - fixed lock conversion deadlock (611892)
74 - fixed the following of symlinks (614622)
75 * recovery: the server can remove locks from a client that dies, other
76 clients can make progress
77 * more extN patch fixes
78 * compile-time configurable ptlrpc buffer allocations
80 - collaborative read cache document
81 - Lustre Lite Performance CDR document-in-progress
83 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
86 - fix typo in patch-2.4.18
88 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
91 - clear ptlrpc request each time in handle_incoming_request()
92 - unlink of files now destroys the object on the OST
94 2002-09-19 Peter Braam <braam@clusterfs.com>
96 * add hard link support
97 * change obdfile creation method
98 * kernel patch changed
100 2002-09-19 Peter Braam <braam@clusterfs.com>
103 - stack overflow bug in extN fixed
105 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
107 * documentation updates
108 - add man pages for config tools
109 - update tests/README to describe testing with new config tools
110 - finish metadata API descriptions
111 * bug fixes and cleanups
112 - statfs workaround for 16TB limit
113 - LOV stripe allocation improved, can stripe on subset of OSTs
114 - LOV file size/IO offset was wrong for files > 4GB in size
115 - object EA data was being dropped, caused files to be unreadable
116 - memory overflow with non-LOV OST caused memory corruption
117 - fixed regression tests to work with new config tools, obdfilter
118 - fixed bug when directory size became larger than 1 block
119 - fixed bug (for single client case) when PWD was deleted
120 - invalidate local directory pages when doing intent-based ops
121 - avoid LDLM oops when lock callback contained bad data
123 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
125 * documentation updates
126 * bug fixes and cleanups
127 - configuration tools
130 - 64-bit compile warnings
131 - 64-bit internal statfs data
133 * test_brw on persistent OST devices
137 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
139 * documentation updates
140 * bug fixes and cleanups
141 * configuration tools
143 2002-08-30 Peter J. Braam <braam@clusterfs.com>
146 * many small fixes to 0_5_4
147 * io/network handling
148 * thinkos in MDS operations
150 2002-08-24 Peter J. Braam <braam@clusterfs.com>
153 * crucial basic fixes to 0.5.3
154 * IOR, Iozone work over Elan
157 2002-08-07 Phil Schwan <phil@clusterfs.com>
158 * version 0_5_3, our first alpha
159 * we use the new Portals iovs
160 * documentation updates
161 * bug fixes and cleanups
162 * small changes in the DLM wire protocol
164 2002-07-25 Peter J. Braam <braam@clusterfs.com>
165 * version 0_5_1 with some initial stability,
166 * locking on MD and file I/O.
167 * documentation updates
168 * several bug fixes since 0.5.0
169 * small changes in wire protocol
171 2002-07-18 Phil Schwan <phil@clusterfs.com>
173 * delivered as Lustre Light Alpha
174 * fixed a crash after handling invalid MDS requests
175 * fixed directory pages for architectures with non-4k pages sizes
177 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
178 * release version v0_4_4
179 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
181 * Fixes a number of interruption problems with OST operations.
182 * Update documentation for portals header changes
183 * Move all wire protocol structs/defines to lustre_idl.h
184 * Fixes symlink length bug.
185 * Add tcpdump to repository.
187 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
188 * release version v0_4_3
189 * Fixes statfs for inodes on extN.
190 * Fixes bug in runtests which would delete /etc/hosts.
191 * Use 64-bit object IDs wherever possible (not into VFS though)
192 Remove ost_get_info, which is unused by lustre, and out of date.
194 2002-07-03 Peter Braam <braam@clusterfs.com>
195 * release version v0_4_2 Fixes a lookup error (type not passed)
196 * move forward to head of Portals
197 * move forward to latest Lustre kernel
199 2002-06-25 Peter Braam <braam@clusterfs.com>
200 * release version v0_4_1. Hopefully stable on single node use.