3 - Fully reactivate OST imports after reconnection (512, others)
4 - Make sure client sees our -ENOTCONN from mds_handle (513 - partial)
5 - More graceful error handling for truncating on dead OST (515)
6 - Don't error out unless we're actually accessing dead stripes (474)
7 - Fix garbage sizes when stripes are missing (410)
8 - LRU counters were broken, causing constant lock purge (433, 432)
9 - garbage on read from stripes with failed OSTs (441)
10 - mark OSCs as active before reconnecting during recovery (438)
11 - lov_enqueue and lov_cancel need to handle inactive OSTs (403)
12 - lfind did not preserve OST order in output (443)
13 - symlinks cause hung clients, incorrect data (439)
14 - stop dereferencing request after dropping refcount (457)
15 - don't LASSERT(spin_is_locked) on non-SMP (455)
16 - fixes for many rename() bugs
17 - fstat didn't correctly synchronize attributes (399)
18 - server must handle lock cancellation during blocking AST prep (487)
19 - bulk descriptors were free()d too soon (511)
20 - fix paths in lconf, which would load incorrect modules (451, 507)
21 - fix confusing lconf 'host not found' error message (386)
22 - fix lock order deadlock on OST (O/R i_sem before journal ops, 478)
24 2002-12-02 Andreas Dilger <adilger@clusterfs.com>
27 - fix many simultaneous client startup (392)
28 - fix dentry->d_it clobbering
29 - credentials weren't being shipped for readdir/getattr operations
30 - remove invalid assertions triggered during some concurrent MD
32 - proper Lustre versions added (336, 389)
33 - fix memory leak for create error case (398)
34 - fix LOV locking bug that would get cli/srv out of sync
35 - fix echo client over LOV (409)
36 - fix dbench 2, extN refcount problem (170, 258, 356, 418)
37 - fix double-O_EXCL intent crash (424)
38 - avoid sending multiple lock CANCELs (352)
40 - MDS can do multi-client recovery (modulo bugs in new code)
42 - many updates, edits, cleanups
44 2002-11-18 Phil Schwan <phil@clusterfs.com>
47 - fix null d_it dereference (346)
48 - fix full OST/dbench hang (333)
49 - fix permission problem with file removal (286)
50 - fix removal of OSCs from LOV when they fail
51 - fix NULL deref during bulk timeout (214)
52 - fix problems related to multiple filesystems on one MDS (241)
53 - fixed serious subtle metadata locking bugs
54 - free locks on clients when inodes are removed due to memory
56 - fix inode pointer in lock data (285)
57 - partial support for multiple MDS on a single host (241)
58 - data locks weren't cancelled at clear_inode time (290, 311)
59 - intent locks could lead to unbounded lock growth (205)
60 - added a maximum lock count, an LRU list, and a flusher
61 - fix multiple rename (365)
62 - properly abstracted the echo client
63 - OSC locked 1 byte too many; fixed
64 - rewrote brw callback code:
65 - fixed recovery bugs related to LOVs (306)
66 - fixed too-many-pages-in-one-write crash (191)
67 - fixed (again) crash in sync_io_timeout (214)
68 - probably fixed callback-related race (385)
70 - Add capability to MDS protocol
71 - LDLM cancellations and callbacks on different portals
73 2002-10-28 Andreas Dilger <adilger@clusterfs.com>
76 - limit client IOV size to PTL_MD_MAX_IOV (611336, 191)
77 - defer open object destruction to close time (601981, 138)
78 - open/close OST file handle in obdo (OBD_MD_FLHANDLE) (601981, 138)
79 - move LDLM_ENQUEUE/CONVERT back to MDS portal (625069)
80 - abstract ll_lookup2, fix ll_revalidate2 to use abstraction (256)
81 - don't call obd_setattr in ll_file_release for destroyed objects
82 * protocol change to lustre_msg: move |version| and add |flags|
83 * protocol change to osc_punch: "start" in "o_size", "end" in "o_blocks"
84 * lock replay: for LDLM_FL_REPLAY trust client to do right thing
85 * added replay of create, unlink, link and rename operations during
86 MDS failover; recovery should be much more robust now
87 * remove failed OSCs from LOVs (only lov_create uses this so far)
88 * the lustre-HOWTO was brought (more) up to date (582544)
90 2002-10-23 Phil Schwan <phil@clusterfs.com>
93 - in-use dentries weren't being reused properly (617851)
94 - prevent multiple LDLM setup (599178)
95 - fix LOV size calculations for truncate (617853)
96 - fix client handling of MDS intent errors (POSIX)
97 - fix permission bug in lovstripe.c test (624321)
98 - fix MDS thread deadlock - move LDLM handler to DLM portal (625069)
99 - truncate past end of file could corrupt data
100 - proper cleanup after timeouts, crashes, etc (592524, 550815)
101 - a race in recovery could return ETIMEDOUT to apps (623947)
102 - building outside the source directory was fixed
103 * the lustre-HOWTO was brought (more) up to date (582544)
104 * major progress was made on recovery functionality
106 2002-10-10 Phil Schwan <phil@clusterfs.com>
109 - recovery deadlock fix
110 - rm -rf causes LBUG fix (617817)
111 - file open by multiple tasks fix (618962)
112 - directory permissions bugs (602707 and 620007)
113 - journal_stop fixed with locking (611313)
114 - O_APPEND failures resolved (618273, perhaps 614459)
115 - lconf PATH fix (619770)
116 - IA64 build fix (621450)
117 - RPC buffer sizes scale with amount of memory
119 2002-10-01 Phil Schwan <phil@clusterfs.com>
122 - locks would be cancelled without throwing away data pages,
123 resulting in inconsistent data (605627)
124 - inode attributes were not always being refreshed (605627, 612449)
125 - lconf now continues to cleanup after lctl reports an error
126 - MDS now enforces user permissions (602707)
127 - lprocfs cleanup fixed, but not yet enabled (614157)
128 - fixed infinite server hang, should a client not respond to an AST
129 - avoid going into recovery if user calls readlink() with a buffer
130 that's too small (613941)
131 - AST RPCs no longer require replies (614867) -- this may be changed
132 - don't crash server if client sends an IOV that's too big (611336)
133 - fixed lock conversion deadlock (611892)
134 - fixed the following of symlinks (614622)
135 * recovery: the server can remove locks from a client that dies, other
136 clients can make progress
137 * more extN patch fixes
138 * compile-time configurable ptlrpc buffer allocations
140 - collaborative read cache document
141 - Lustre Lite Performance CDR document-in-progress
143 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
146 - fix typo in patch-2.4.18
148 2002-09-20 Andreas Dilger <adilger@clusterfs.com>
151 - clear ptlrpc request each time in handle_incoming_request()
152 - unlink of files now destroys the object on the OST
154 2002-09-19 Peter Braam <braam@clusterfs.com>
156 * add hard link support
157 * change obdfile creation method
158 * kernel patch changed
160 2002-09-19 Peter Braam <braam@clusterfs.com>
163 - stack overflow bug in extN fixed
165 2002-09-18 Andreas Dilger <adilger@clusterfs.com>
167 * documentation updates
168 - add man pages for config tools
169 - update tests/README to describe testing with new config tools
170 - finish metadata API descriptions
171 * bug fixes and cleanups
172 - statfs workaround for 16TB limit
173 - LOV stripe allocation improved, can stripe on subset of OSTs
174 - LOV file size/IO offset was wrong for files > 4GB in size
175 - object EA data was being dropped, caused files to be unreadable
176 - memory overflow with non-LOV OST caused memory corruption
177 - fixed regression tests to work with new config tools, obdfilter
178 - fixed bug when directory size became larger than 1 block
179 - fixed bug (for single client case) when PWD was deleted
180 - invalidate local directory pages when doing intent-based ops
181 - avoid LDLM oops when lock callback contained bad data
183 2002-09-09 Andreas Dilger <adilger@clusterfs.com>
185 * documentation updates
186 * bug fixes and cleanups
187 - configuration tools
190 - 64-bit compile warnings
191 - 64-bit internal statfs data
193 * test_brw on persistent OST devices
197 2002-09-04 Andreas Dilger <adilger@clusterfs.com>
199 * documentation updates
200 * bug fixes and cleanups
201 * configuration tools
203 2002-08-30 Peter J. Braam <braam@clusterfs.com>
206 * many small fixes to 0_5_4
207 * io/network handling
208 * thinkos in MDS operations
210 2002-08-24 Peter J. Braam <braam@clusterfs.com>
213 * crucial basic fixes to 0.5.3
214 * IOR, Iozone work over Elan
217 2002-08-07 Phil Schwan <phil@clusterfs.com>
218 * version 0_5_3, our first alpha
219 * we use the new Portals iovs
220 * documentation updates
221 * bug fixes and cleanups
222 * small changes in the DLM wire protocol
224 2002-07-25 Peter J. Braam <braam@clusterfs.com>
225 * version 0_5_1 with some initial stability,
226 * locking on MD and file I/O.
227 * documentation updates
228 * several bug fixes since 0.5.0
229 * small changes in wire protocol
231 2002-07-18 Phil Schwan <phil@clusterfs.com>
233 * delivered as Lustre Light Alpha
234 * fixed a crash after handling invalid MDS requests
235 * fixed directory pages for architectures with non-4k pages sizes
237 2002-07-11 Andreas Dilger <adilger@clusterfs.com>
238 * release version v0_4_4
239 * Moves TCP acceptor to be on port 2432 (unused Coda port) instead
241 * Fixes a number of interruption problems with OST operations.
242 * Update documentation for portals header changes
243 * Move all wire protocol structs/defines to lustre_idl.h
244 * Fixes symlink length bug.
245 * Add tcpdump to repository.
247 2002-07-05 Andreas Dilger <adilger@clusterfs.com>
248 * release version v0_4_3
249 * Fixes statfs for inodes on extN.
250 * Fixes bug in runtests which would delete /etc/hosts.
251 * Use 64-bit object IDs wherever possible (not into VFS though)
252 Remove ost_get_info, which is unused by lustre, and out of date.
254 2002-07-03 Peter Braam <braam@clusterfs.com>
255 * release version v0_4_2 Fixes a lookup error (type not passed)
256 * move forward to head of Portals
257 * move forward to latest Lustre kernel
259 2002-06-25 Peter Braam <braam@clusterfs.com>
260 * release version v0_4_1. Hopefully stable on single node use.