4 - limit Lustre IOVs to PTL_MD_MAX_IOV (611336)
5 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
6 * protocol change to lustre_msg: move |version| and add |flags|
7 * added replay of create, unlink, link and rename operations during
8 MDS failover; recovery should be much more robust now
9 * remove failed OSCs from LOVs
11 2002-10-23 Phil Schwan <phil@clusterfs.com>
14 - in-use dentries weren't being reused properly (617851)
15 - prevent multiple LDLM setup (599178)
16 - fix LOV size calculations for truncate (617853)
17 - fix client handling of MDS intent errors (POSIX)
18 - fix permission bug in lovstripe.c test (624321)
19 - fix MDS thread deadlock (625069)
20 - truncate past end of file could corrupt data
21 - proper cleanup after timeouts, crashes, etc (592524, 550815)
22 - a race in recovery could return ETIMEDOUT to apps (623947)
23 - building outside the source directory was fixed
24 * the howto was brought (more) up to date (582544)
25 * major progress was made on recovery functionality
27 2002-10-10 Phil Schwan <phil@clusterfs.com>
30 - recovery deadlock fix
31 - rm -rf causes LBUG fix (617817)
32 - file open by multiple tasks fix (618962)
33 - directory permissions bugs (602707 and 620007)
34 - journal_stop fixed with locking (611313)
35 - O_APPEND failures resolved (618273, perhaps 614459)
36 - lconf PATH fix (619770)
37 - IA64 build fix (621450)
38 - RPC buffer sizes scale with amount of memory
40 2002-10-01 Phil Schwan <phil@clusterfs.com>
43 - locks would be cancelled without throwing away data pages,
44 resulting in inconsistent data (605627)
45 - inode attributes were not always being refreshed (605627, 612449)
46 - lconf now continues to cleanup after lctl reports an error
47 - MDS now enforces user permissions (602707)
48 - lprocfs cleanup fixed, but not yet enabled (614157)
49 - fixed infinite server hang, should a client not respond to an AST
50 - avoid going into recovery if user calls readlink() with a buffer
51 that's too small (613941)
52 - AST RPCs no longer require replies (614867) -- this may be changed
53 - don't crash server if client sends an IOV that's too big (611336)
54 - fixed lock conversion deadlock (611892)
55 - fixed the following of symlinks (614622)
56 * recovery: the server can remove locks from a client that dies, other
57 clients can make progress
58 * more extN patch fixes
59 * compile-time configurable ptlrpc buffer allocations
61 - collaborative read cache document
62 - Lustre Lite Performance CDR document-in-progress
64 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
67 - fix typo in patch-2.4.18
69 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
72 - clear ptlrpc request each time in handle_incoming_request()
73 - unlink of files now destroys the object on the OST
75 2002-09-19 Peter Braam <braam@clusterfs.com>
77 * add hard link support
78 * change obdfile creation method
79 * kernel patch changed
81 2002-09-19 Peter Braam <braam@clusterfs.com>
84 - stack overflow bug in extN fixed
86 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
88 * documentation updates
89 - add man pages for config tools
90 - update tests/README to describe testing with new config tools
91 - finish metadata API descriptions
92 * bug fixes and cleanups
93 - statfs workaround for 16TB limit
94 - LOV stripe allocation improved, can stripe on subset of OSTs
95 - LOV file size/IO offset was wrong for files > 4GB in size
96 - object EA data was being dropped, caused files to be unreadable
97 - memory overflow with non-LOV OST caused memory corruption
98 - fixed regression tests to work with new config tools, obdfilter
99 - fixed bug when directory size became larger than 1 block
100 - fixed bug (for single client case) when PWD was deleted
101 - invalidate local directory pages when doing intent-based ops
102 - avoid LDLM oops when lock callback contained bad data
104 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
106 * documentation updates
107 * bug fixes and cleanups
108 - configuration tools
111 - 64-bit compile warnings
112 - 64-bit internal statfs data
114 * test_brw on persistent OST devices
118 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
120 * documentation updates
121 * bug fixes and cleanups
122 * configuration tools
124 2002-08-30 Peter J. Braam <braam@clusterfs.com>
127 * many small fixes to 0_5_4
128 * io/network handling
129 * thinkos in MDS operations
131 2002-08-24 Peter J. Braam <braam@clusterfs.com>
134 * crucial basic fixes to 0.5.3
135 * IOR, Iozone work over Elan
138 2002-08-07 Phil Schwan <phil@clusterfs.com>
139 * version 0_5_3, our first alpha
140 * we use the new Portals iovs
141 * documentation updates
142 * bug fixes and cleanups
143 * small changes in the DLM wire protocol
145 2002-07-25 Peter J. Braam <braam@clusterfs.com>
146 * version 0_5_1 with some initial stability,
147 * locking on MD and file I/O.
148 * documentation updates
149 * several bug fixes since 0.5.0
150 * small changes in wire protocol
152 2002-07-18 Phil Schwan <phil@clusterfs.com>
154 * delivered as Lustre Light Alpha
155 * fixed a crash after handling invalid MDS requests
156 * fixed directory pages for architectures with non-4k pages sizes
158 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
159 * release version v0_4_4
160 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
162 * Fixes a number of interruption problems with OST operations.
163 * Update documentation for portals header changes
164 * Move all wire protocol structs/defines to lustre_idl.h
165 * Fixes symlink length bug.
166 * Add tcpdump to repository.
168 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
169 * release version v0_4_3
170 * Fixes statfs for inodes on extN.
171 * Fixes bug in runtests which would delete /etc/hosts.
172 * Use 64-bit object IDs wherever possible (not into VFS though)
173 Remove ost_get_info, which is unused by lustre, and out of date.
175 2002-07-03 Peter Braam <braam@clusterfs.com>
176 * release version v0_4_2 Fixes a lookup error (type not passed)
177 * move forward to head of Portals
178 * move forward to latest Lustre kernel
180 2002-06-25 Peter Braam <braam@clusterfs.com>
181 * release version v0_4_1. Hopefully stable on single node use.