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)
12 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
15 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
16 - defer open object destruction to close time (601981, 138)
17 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
18 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
19 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
20 - don't call obd_setattr in ll_file_release for destroyed objects
21 * protocol change to lustre_msg: move |version| and add |flags|
22 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
23 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
24 * added replay of create, unlink, link and rename operations during
25 MDS failover; recovery should be much more robust now
26 * remove failed OSCs from LOVs (only lov_create uses this so far)
27 * the lustre-HOWTO was brought (more) up to date (582544)
29 2002-10-23 Phil Schwan <phil@clusterfs.com>
32 - in-use dentries weren't being reused properly (617851)
33 - prevent multiple LDLM setup (599178)
34 - fix LOV size calculations for truncate (617853)
35 - fix client handling of MDS intent errors (POSIX)
36 - fix permission bug in lovstripe.c test (624321)
37 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
38 - truncate past end of file could corrupt data
39 - proper cleanup after timeouts, crashes, etc (592524, 550815)
40 - a race in recovery could return ETIMEDOUT to apps (623947)
41 - building outside the source directory was fixed
42 * the lustre-HOWTO was brought (more) up to date (582544)
43 * major progress was made on recovery functionality
45 2002-10-10 Phil Schwan <phil@clusterfs.com>
48 - recovery deadlock fix
49 - rm -rf causes LBUG fix (617817)
50 - file open by multiple tasks fix (618962)
51 - directory permissions bugs (602707 and 620007)
52 - journal_stop fixed with locking (611313)
53 - O_APPEND failures resolved (618273, perhaps 614459)
54 - lconf PATH fix (619770)
55 - IA64 build fix (621450)
56 - RPC buffer sizes scale with amount of memory
58 2002-10-01 Phil Schwan <phil@clusterfs.com>
61 - locks would be cancelled without throwing away data pages,
62 resulting in inconsistent data (605627)
63 - inode attributes were not always being refreshed (605627, 612449)
64 - lconf now continues to cleanup after lctl reports an error
65 - MDS now enforces user permissions (602707)
66 - lprocfs cleanup fixed, but not yet enabled (614157)
67 - fixed infinite server hang, should a client not respond to an AST
68 - avoid going into recovery if user calls readlink() with a buffer
69 that's too small (613941)
70 - AST RPCs no longer require replies (614867) -- this may be changed
71 - don't crash server if client sends an IOV that's too big (611336)
72 - fixed lock conversion deadlock (611892)
73 - fixed the following of symlinks (614622)
74 * recovery: the server can remove locks from a client that dies, other
75 clients can make progress
76 * more extN patch fixes
77 * compile-time configurable ptlrpc buffer allocations
79 - collaborative read cache document
80 - Lustre Lite Performance CDR document-in-progress
82 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
85 - fix typo in patch-2.4.18
87 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
90 - clear ptlrpc request each time in handle_incoming_request()
91 - unlink of files now destroys the object on the OST
93 2002-09-19 Peter Braam <braam@clusterfs.com>
95 * add hard link support
96 * change obdfile creation method
97 * kernel patch changed
99 2002-09-19 Peter Braam <braam@clusterfs.com>
102 - stack overflow bug in extN fixed
104 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
106 * documentation updates
107 - add man pages for config tools
108 - update tests/README to describe testing with new config tools
109 - finish metadata API descriptions
110 * bug fixes and cleanups
111 - statfs workaround for 16TB limit
112 - LOV stripe allocation improved, can stripe on subset of OSTs
113 - LOV file size/IO offset was wrong for files > 4GB in size
114 - object EA data was being dropped, caused files to be unreadable
115 - memory overflow with non-LOV OST caused memory corruption
116 - fixed regression tests to work with new config tools, obdfilter
117 - fixed bug when directory size became larger than 1 block
118 - fixed bug (for single client case) when PWD was deleted
119 - invalidate local directory pages when doing intent-based ops
120 - avoid LDLM oops when lock callback contained bad data
122 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
124 * documentation updates
125 * bug fixes and cleanups
126 - configuration tools
129 - 64-bit compile warnings
130 - 64-bit internal statfs data
132 * test_brw on persistent OST devices
136 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
138 * documentation updates
139 * bug fixes and cleanups
140 * configuration tools
142 2002-08-30 Peter J. Braam <braam@clusterfs.com>
145 * many small fixes to 0_5_4
146 * io/network handling
147 * thinkos in MDS operations
149 2002-08-24 Peter J. Braam <braam@clusterfs.com>
152 * crucial basic fixes to 0.5.3
153 * IOR, Iozone work over Elan
156 2002-08-07 Phil Schwan <phil@clusterfs.com>
157 * version 0_5_3, our first alpha
158 * we use the new Portals iovs
159 * documentation updates
160 * bug fixes and cleanups
161 * small changes in the DLM wire protocol
163 2002-07-25 Peter J. Braam <braam@clusterfs.com>
164 * version 0_5_1 with some initial stability,
165 * locking on MD and file I/O.
166 * documentation updates
167 * several bug fixes since 0.5.0
168 * small changes in wire protocol
170 2002-07-18 Phil Schwan <phil@clusterfs.com>
172 * delivered as Lustre Light Alpha
173 * fixed a crash after handling invalid MDS requests
174 * fixed directory pages for architectures with non-4k pages sizes
176 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
177 * release version v0_4_4
178 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
180 * Fixes a number of interruption problems with OST operations.
181 * Update documentation for portals header changes
182 * Move all wire protocol structs/defines to lustre_idl.h
183 * Fixes symlink length bug.
184 * Add tcpdump to repository.
186 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
187 * release version v0_4_3
188 * Fixes statfs for inodes on extN.
189 * Fixes bug in runtests which would delete /etc/hosts.
190 * Use 64-bit object IDs wherever possible (not into VFS though)
191 Remove ost_get_info, which is unused by lustre, and out of date.
193 2002-07-03 Peter Braam <braam@clusterfs.com>
194 * release version v0_4_2 Fixes a lookup error (type not passed)
195 * move forward to head of Portals
196 * move forward to latest Lustre kernel
198 2002-06-25 Peter Braam <braam@clusterfs.com>
199 * release version v0_4_1. Hopefully stable on single node use.