1 2002-10-23 Phil Schwan <phil@clusterfs.com>
4 - in-use dentries weren't being reused properly (617851)
5 - prevent multiple LDLM setup (599178)
6 - fix LOV size calculations for truncate (617853)
7 - fix client handling of MDS intent errors (POSIX)
8 - fix permission bug in lovstripe.c test (624321)
9 - fix MDS thread deadlock (625069)
10 - truncate past end of file could corrupt data
11 - proper cleanup after timeouts, crashes, etc (592524, 550815)
12 - a race in recovery could return ETIMEDOUT to apps (623947)
13 - building outside the source directory was fixed
14 * the howto was brought (more) up to date (582544)
15 * major progress was made on recovery functionality
17 2002-10-10 Phil Schwan <phil@clusterfs.com>
20 - recovery deadlock fix
21 - rm -rf causes LBUG fix (617817)
22 - file open by multiple tasks fix (618962)
23 - directory permissions bugs (602707 and 620007)
24 - journal_stop fixed with locking (611313)
25 - O_APPEND failures resolved (618273, perhaps 614459)
26 - lconf PATH fix (619770)
27 - IA64 build fix (621450)
28 - RPC buffer sizes scale with amount of memory
30 2002-10-01 Phil Schwan <phil@clusterfs.com>
33 - locks would be cancelled without throwing away data pages,
34 resulting in inconsistent data (605627)
35 - inode attributes were not always being refreshed (605627, 612449)
36 - lconf now continues to cleanup after lctl reports an error
37 - MDS now enforces user permissions (602707)
38 - lprocfs cleanup fixed, but not yet enabled (614157)
39 - fixed infinite server hang, should a client not respond to an AST
40 - avoid going into recovery if user calls readlink() with a buffer
41 that's too small (613941)
42 - AST RPCs no longer require replies (614867) -- this may be changed
43 - don't crash server if client sends an IOV that's too big (611336)
44 - fixed lock conversion deadlock (611892)
45 - fixed the following of symlinks (614622)
46 * recovery: the server can remove locks from a client that dies, other
47 clients can make progress
48 * more extN patch fixes
49 * compile-time configurable ptlrpc buffer allocations
51 - collaborative read cache document
52 - Lustre Lite Performance CDR document-in-progress
54 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
57 - fix typo in patch-2.4.18
59 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
62 - clear ptlrpc request each time in handle_incoming_request()
63 - unlink of files now destroys the object on the OST
65 2002-09-19 Peter Braam <braam@clusterfs.com>
67 * add hard link support
68 * change obdfile creation method
69 * kernel patch changed
71 2002-09-19 Peter Braam <braam@clusterfs.com>
74 - stack overflow bug in extN fixed
76 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
78 * documentation updates
79 - add man pages for config tools
80 - update tests/README to describe testing with new config tools
81 - finish metadata API descriptions
82 * bug fixes and cleanups
83 - statfs workaround for 16TB limit
84 - LOV stripe allocation improved, can stripe on subset of OSTs
85 - LOV file size/IO offset was wrong for files > 4GB in size
86 - object EA data was being dropped, caused files to be unreadable
87 - memory overflow with non-LOV OST caused memory corruption
88 - fixed regression tests to work with new config tools, obdfilter
89 - fixed bug when directory size became larger than 1 block
90 - fixed bug (for single client case) when PWD was deleted
91 - invalidate local directory pages when doing intent-based ops
92 - avoid LDLM oops when lock callback contained bad data
94 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
96 * documentation updates
97 * bug fixes and cleanups
101 - 64-bit compile warnings
102 - 64-bit internal statfs data
104 * test_brw on persistent OST devices
108 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
110 * documentation updates
111 * bug fixes and cleanups
112 * configuration tools
114 2002-08-30 Peter J. Braam <braam@clusterfs.com>
117 * many small fixes to 0_5_4
118 * io/network handling
119 * thinkos in MDS operations
121 2002-08-24 Peter J. Braam <braam@clusterfs.com>
124 * crucial basic fixes to 0.5.3
125 * IOR, Iozone work over Elan
128 2002-08-07 Phil Schwan <phil@clusterfs.com>
129 * version 0_5_3, our first alpha
130 * we use the new Portals iovs
131 * documentation updates
132 * bug fixes and cleanups
133 * small changes in the DLM wire protocol
135 2002-07-25 Peter J. Braam <braam@clusterfs.com>
136 * version 0_5_1 with some initial stability,
137 * locking on MD and file I/O.
138 * documentation updates
139 * several bug fixes since 0.5.0
140 * small changes in wire protocol
142 2002-07-18 Phil Schwan <phil@clusterfs.com>
144 * delivered as Lustre Light Alpha
145 * fixed a crash after handling invalid MDS requests
146 * fixed directory pages for architectures with non-4k pages sizes
148 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
149 * release version v0_4_4
150 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
152 * Fixes a number of interruption problems with OST operations.
153 * Update documentation for portals header changes
154 * Move all wire protocol structs/defines to lustre_idl.h
155 * Fixes symlink length bug.
156 * Add tcpdump to repository.
158 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
159 * release version v0_4_3
160 * Fixes statfs for inodes on extN.
161 * Fixes bug in runtests which would delete /etc/hosts.
162 * Use 64-bit object IDs wherever possible (not into VFS though)
163 Remove ost_get_info, which is unused by lustre, and out of date.
165 2002-07-03 Peter Braam <braam@clusterfs.com>
166 * release version v0_4_2 Fixes a lookup error (type not passed)
167 * move forward to head of Portals
168 * move forward to latest Lustre kernel
170 2002-06-25 Peter Braam <braam@clusterfs.com>
171 * release version v0_4_1. Hopefully stable on single node use.