1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
|
From 50318713486e79d9258cf22e656caff402256dde Mon Sep 17 00:00:00 2001
From: Mohit Agrawal <moagrawal@redhat.com>
Date: Sun, 20 Oct 2019 22:01:01 +0530
Subject: [PATCH 389/449] posix: Avoid diskpace error in case of overwriting
the data
Problem: Sometime fops like posix_writev, posix_fallocate, posix_zerofile
failed and throw error ENOSPC if storage.reserve threshold limit
has reached even fops is overwriting the data
Solution: Retry the fops in case of overwrite if diskspace check
is failed
> Credits: kinsu <vpolakis@gmail.com>
> Change-Id: I987d73bcf47ed1bb27878df40c39751296e95fe8
> Updates: #745
> Signed-off-by: Mohit Agrawal <moagrawa@redhat.com>
> (Cherry pick from commit ca3e5905ac02fb9c373ac3de10b44f061d04cd6f)
> (Reviewed on upstream link https://review.gluster.org/#/c/glusterfs/+/23572/)
Change-Id: I987d73bcf47ed1bb27878df40c39751296e95fe8
BUG: 1787331
Signed-off-by: Mohit Agrawal <moagrawa@redhat.com>
Reviewed-on: https://code.engineering.redhat.com/gerrit/202307
Tested-by: RHGS Build Bot <nigelb@redhat.com>
Reviewed-by: Sunil Kumar Heggodu Gopala Acharya <sheggodu@redhat.com>
---
tests/bugs/posix/bug-1651445.t | 1 +
xlators/storage/posix/src/posix-entry-ops.c | 1 -
xlators/storage/posix/src/posix-inode-fd-ops.c | 141 ++++++++++++++++++++++---
3 files changed, 126 insertions(+), 17 deletions(-)
diff --git a/tests/bugs/posix/bug-1651445.t b/tests/bugs/posix/bug-1651445.t
index 5248d47..4d08b69 100644
--- a/tests/bugs/posix/bug-1651445.t
+++ b/tests/bugs/posix/bug-1651445.t
@@ -33,6 +33,7 @@ sleep 5
# setup_lvm create lvm partition of 150M and 40M are reserve so after
# consuming more than 110M next dd should fail
TEST ! dd if=/dev/zero of=$M0/c bs=5M count=1
+TEST dd if=/dev/urandom of=$M0/a bs=1022 count=1 oflag=seek_bytes,sync seek=102 conv=notrunc
rm -rf $M0/*
diff --git a/xlators/storage/posix/src/posix-entry-ops.c b/xlators/storage/posix/src/posix-entry-ops.c
index 283b305..bea0bbf 100644
--- a/xlators/storage/posix/src/posix-entry-ops.c
+++ b/xlators/storage/posix/src/posix-entry-ops.c
@@ -1634,7 +1634,6 @@ posix_rename(call_frame_t *frame, xlator_t *this, loc_t *oldloc, loc_t *newloc,
priv = this->private;
VALIDATE_OR_GOTO(priv, out);
- DISK_SPACE_CHECK_AND_GOTO(frame, priv, xdata, op_ret, op_errno, out);
SET_FS_ID(frame->root->uid, frame->root->gid);
MAKE_ENTRY_HANDLE(real_oldpath, par_oldpath, this, oldloc, NULL);
diff --git a/xlators/storage/posix/src/posix-inode-fd-ops.c b/xlators/storage/posix/src/posix-inode-fd-ops.c
index a2a518f..bcce06e 100644
--- a/xlators/storage/posix/src/posix-inode-fd-ops.c
+++ b/xlators/storage/posix/src/posix-inode-fd-ops.c
@@ -692,6 +692,10 @@ posix_do_fallocate(call_frame_t *frame, xlator_t *this, fd_t *fd, int32_t flags,
gf_boolean_t locked = _gf_false;
posix_inode_ctx_t *ctx = NULL;
struct posix_private *priv = NULL;
+ gf_boolean_t check_space_error = _gf_false;
+ struct stat statbuf = {
+ 0,
+ };
DECLARE_OLD_FS_ID_VAR;
@@ -711,7 +715,10 @@ posix_do_fallocate(call_frame_t *frame, xlator_t *this, fd_t *fd, int32_t flags,
if (priv->disk_reserve)
posix_disk_space_check(this);
- DISK_SPACE_CHECK_AND_GOTO(frame, priv, xdata, ret, ret, out);
+ DISK_SPACE_CHECK_AND_GOTO(frame, priv, xdata, ret, ret, unlock);
+
+overwrite:
+ check_space_error = _gf_true;
ret = posix_fd_ctx_get(fd, this, &pfd, &op_errno);
if (ret < 0) {
@@ -735,7 +742,7 @@ posix_do_fallocate(call_frame_t *frame, xlator_t *this, fd_t *fd, int32_t flags,
ret = -errno;
gf_msg(this->name, GF_LOG_ERROR, errno, P_MSG_FSTAT_FAILED,
"fallocate (fstat) failed on fd=%p", fd);
- goto out;
+ goto unlock;
}
if (xdata) {
@@ -745,7 +752,7 @@ posix_do_fallocate(call_frame_t *frame, xlator_t *this, fd_t *fd, int32_t flags,
gf_msg(this->name, GF_LOG_ERROR, 0, 0,
"file state check failed, fd %p", fd);
ret = -EIO;
- goto out;
+ goto unlock;
}
}
@@ -756,7 +763,7 @@ posix_do_fallocate(call_frame_t *frame, xlator_t *this, fd_t *fd, int32_t flags,
"fallocate failed on %s offset: %jd, "
"len:%zu, flags: %d",
uuid_utoa(fd->inode->gfid), offset, len, flags);
- goto out;
+ goto unlock;
}
ret = posix_fdstat(this, fd->inode, pfd->fd, statpost);
@@ -764,16 +771,47 @@ posix_do_fallocate(call_frame_t *frame, xlator_t *this, fd_t *fd, int32_t flags,
ret = -errno;
gf_msg(this->name, GF_LOG_ERROR, errno, P_MSG_FSTAT_FAILED,
"fallocate (fstat) failed on fd=%p", fd);
- goto out;
+ goto unlock;
}
posix_set_ctime(frame, this, NULL, pfd->fd, fd->inode, statpost);
-out:
+unlock:
if (locked) {
pthread_mutex_unlock(&ctx->write_atomic_lock);
locked = _gf_false;
}
+
+ if (op_errno == ENOSPC && priv->disk_space_full && !check_space_error) {
+#ifdef FALLOC_FL_KEEP_SIZE
+ if (flags & FALLOC_FL_KEEP_SIZE) {
+ goto overwrite;
+ }
+#endif
+ ret = posix_fd_ctx_get(fd, this, &pfd, &op_errno);
+ if (ret < 0) {
+ gf_msg(this->name, GF_LOG_WARNING, ret, P_MSG_PFD_NULL,
+ "pfd is NULL from fd=%p", fd);
+ goto out;
+ }
+
+ if (sys_fstat(pfd->fd, &statbuf) < 0) {
+ gf_msg(this->name, GF_LOG_WARNING, op_errno, P_MSG_FILE_OP_FAILED,
+ "%d", pfd->fd);
+ goto out;
+ }
+
+ if (offset + len <= statbuf.st_size) {
+ gf_msg_debug(this->name, 0,
+ "io vector size will not"
+ " change disk size so allow overwrite for"
+ " fd %d",
+ pfd->fd);
+ goto overwrite;
+ }
+ }
+
+out:
SET_TO_OLD_FS_ID();
if (ret == ENOSPC)
ret = -ENOSPC;
@@ -1083,25 +1121,57 @@ posix_zerofill(call_frame_t *frame, xlator_t *this, fd_t *fd, off_t offset,
int op_ret = -1;
int op_errno = EINVAL;
dict_t *rsp_xdata = NULL;
+ gf_boolean_t check_space_error = _gf_false;
+ struct posix_fd *pfd = NULL;
+ struct stat statbuf = {
+ 0,
+ };
- VALIDATE_OR_GOTO(frame, out);
- VALIDATE_OR_GOTO(this, out);
+ VALIDATE_OR_GOTO(frame, unwind);
+ VALIDATE_OR_GOTO(this, unwind);
priv = this->private;
DISK_SPACE_CHECK_AND_GOTO(frame, priv, xdata, op_ret, op_errno, out);
+overwrite:
+ check_space_error = _gf_true;
ret = posix_do_zerofill(frame, this, fd, offset, len, &statpre, &statpost,
xdata, &rsp_xdata);
if (ret < 0) {
op_ret = -1;
op_errno = -ret;
- goto out;
+ goto unwind;
}
STACK_UNWIND_STRICT(zerofill, frame, 0, 0, &statpre, &statpost, rsp_xdata);
return 0;
out:
+ if (op_errno == ENOSPC && priv->disk_space_full && !check_space_error) {
+ ret = posix_fd_ctx_get(fd, this, &pfd, &op_errno);
+ if (ret < 0) {
+ gf_msg(this->name, GF_LOG_WARNING, ret, P_MSG_PFD_NULL,
+ "pfd is NULL from fd=%p", fd);
+ goto out;
+ }
+
+ if (sys_fstat(pfd->fd, &statbuf) < 0) {
+ gf_msg(this->name, GF_LOG_WARNING, op_errno, P_MSG_FILE_OP_FAILED,
+ "%d", pfd->fd);
+ goto out;
+ }
+
+ if (offset + len <= statbuf.st_size) {
+ gf_msg_debug(this->name, 0,
+ "io vector size will not"
+ " change disk size so allow overwrite for"
+ " fd %d",
+ pfd->fd);
+ goto overwrite;
+ }
+ }
+
+unwind:
STACK_UNWIND_STRICT(zerofill, frame, op_ret, op_errno, NULL, NULL,
rsp_xdata);
return 0;
@@ -1857,19 +1927,28 @@ posix_writev(call_frame_t *frame, xlator_t *this, fd_t *fd,
gf_boolean_t write_append = _gf_false;
gf_boolean_t update_atomic = _gf_false;
posix_inode_ctx_t *ctx = NULL;
+ gf_boolean_t check_space_error = _gf_false;
+ struct stat statbuf = {
+ 0,
+ };
+ int totlen = 0;
+ int idx = 0;
- VALIDATE_OR_GOTO(frame, out);
- VALIDATE_OR_GOTO(this, out);
- VALIDATE_OR_GOTO(fd, out);
- VALIDATE_OR_GOTO(fd->inode, out);
- VALIDATE_OR_GOTO(vector, out);
- VALIDATE_OR_GOTO(this->private, out);
+ VALIDATE_OR_GOTO(frame, unwind);
+ VALIDATE_OR_GOTO(this, unwind);
+ VALIDATE_OR_GOTO(fd, unwind);
+ VALIDATE_OR_GOTO(fd->inode, unwind);
+ VALIDATE_OR_GOTO(vector, unwind);
+ VALIDATE_OR_GOTO(this->private, unwind);
priv = this->private;
- VALIDATE_OR_GOTO(priv, out);
+ VALIDATE_OR_GOTO(priv, unwind);
DISK_SPACE_CHECK_AND_GOTO(frame, priv, xdata, op_ret, op_errno, out);
+overwrite:
+
+ check_space_error = _gf_true;
if ((fd->inode->ia_type == IA_IFBLK) || (fd->inode->ia_type == IA_IFCHR)) {
gf_msg(this->name, GF_LOG_ERROR, EINVAL, P_MSG_INVALID_ARGUMENT,
"writev received on a block/char file (%s)",
@@ -2011,6 +2090,36 @@ out:
locked = _gf_false;
}
+ if (op_errno == ENOSPC && priv->disk_space_full && !check_space_error) {
+ ret = posix_fd_ctx_get(fd, this, &pfd, &op_errno);
+ if (ret < 0) {
+ gf_msg(this->name, GF_LOG_WARNING, ret, P_MSG_PFD_NULL,
+ "pfd is NULL from fd=%p", fd);
+ goto unwind;
+ }
+
+ if (sys_fstat(pfd->fd, &statbuf) < 0) {
+ gf_msg(this->name, GF_LOG_WARNING, op_errno, P_MSG_FILE_OP_FAILED,
+ "%d", pfd->fd);
+ goto unwind;
+ }
+
+ for (idx = 0; idx < count; idx++) {
+ totlen = vector[idx].iov_len;
+ }
+
+ if ((offset + totlen <= statbuf.st_size) &&
+ !(statbuf.st_blocks * statbuf.st_blksize < statbuf.st_size)) {
+ gf_msg_debug(this->name, 0,
+ "io vector size will not"
+ " change disk size so allow overwrite for"
+ " fd %d",
+ pfd->fd);
+ goto overwrite;
+ }
+ }
+
+unwind:
STACK_UNWIND_STRICT(writev, frame, op_ret, op_errno, &preop, &postop,
rsp_xdata);
--
1.8.3.1
|
