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
|
From 87d8070f80487322a1736846a78725fd88f8de34 Mon Sep 17 00:00:00 2001
From: Pranith Kumar K <pkarampu@redhat.com>
Date: Tue, 20 Aug 2019 13:27:24 +0530
Subject: [PATCH 291/297] cluster/ec: Mark release only when it is acquired
Problem:
Mount-1 Mount-2
1)Tries to acquire lock on 'dir1' 1)Tries to acquire lock on 'dir1'
2)Lock is granted on brick-0 2)Lock gets EAGAIN on brick-0 and
leads to blocking lock on brick-0
3)Gets a lock-contention 3) Doesn't matter what happens on mount-2
notification, marks lock->release from here on.
to true.
4)New fop comes on 'dir1' which will
be put in frozen list as lock->release
is set to true.
5) Lock acquisition from step-2 fails because
3 bricks went down in 4+2 setup.
Fop on mount-1 which is put in frozen list will hang because no codepath will
move it from frozen list to any other list and the lock will not be retried.
Fix:
Don't set lock->release to true if lock is not acquired at the time of
lock-contention-notification
Upstream-patch: https://review.gluster.org/c/glusterfs/+/23272
fixes: bz#1731896
Change-Id: Ie6630db8735ccf372cc54b873a3a3aed7a6082b7
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>
Reviewed-on: https://code.engineering.redhat.com/gerrit/180870
Tested-by: RHGS Build Bot <nigelb@redhat.com>
Reviewed-by: Ashish Pandey <aspandey@redhat.com>
Reviewed-by: Sunil Kumar Heggodu Gopala Acharya <sheggodu@redhat.com>
---
xlators/cluster/ec/src/ec-common.c | 20 ++++++++++++++++++--
xlators/cluster/ec/src/ec-types.h | 1 +
2 files changed, 19 insertions(+), 2 deletions(-)
diff --git a/xlators/cluster/ec/src/ec-common.c b/xlators/cluster/ec/src/ec-common.c
index 2e59180..5cae37b 100644
--- a/xlators/cluster/ec/src/ec-common.c
+++ b/xlators/cluster/ec/src/ec-common.c
@@ -1867,6 +1867,10 @@ ec_lock_acquired(ec_lock_link_t *link)
LOCK(&lock->loc.inode->lock);
lock->acquired = _gf_true;
+ if (lock->contention) {
+ lock->release = _gf_true;
+ lock->contention = _gf_false;
+ }
ec_lock_update_fd(lock, fop);
ec_lock_wake_shared(lock, &list);
@@ -1892,15 +1896,20 @@ ec_locked(call_frame_t *frame, void *cookie, xlator_t *this, int32_t op_ret,
ec_lock_link_t *link = NULL;
ec_lock_t *lock = NULL;
+ link = fop->data;
+ lock = link->lock;
if (op_ret >= 0) {
- link = fop->data;
- lock = link->lock;
lock->mask = lock->good_mask = fop->good;
lock->healing = 0;
ec_lock_acquired(link);
ec_lock(fop->parent);
} else {
+ LOCK(&lock->loc.inode->lock);
+ {
+ lock->contention = _gf_false;
+ }
+ UNLOCK(&lock->loc.inode->lock);
gf_msg(this->name, GF_LOG_WARNING, op_errno, EC_MSG_PREOP_LOCK_FAILED,
"Failed to complete preop lock");
}
@@ -2547,6 +2556,13 @@ ec_lock_release(ec_t *ec, inode_t *inode)
gf_msg_debug(ec->xl->name, 0, "Releasing inode %p due to lock contention",
inode);
+ if (!lock->acquired) {
+ /* This happens if some bricks already got the lock while inodelk is in
+ * progress. Set release to true after lock is acquired*/
+ lock->contention = _gf_true;
+ goto done;
+ }
+
/* The lock is not marked to be released, so the frozen list should be
* empty. */
GF_ASSERT(list_empty(&lock->frozen));
diff --git a/xlators/cluster/ec/src/ec-types.h b/xlators/cluster/ec/src/ec-types.h
index ea4f6ad..34a9768 100644
--- a/xlators/cluster/ec/src/ec-types.h
+++ b/xlators/cluster/ec/src/ec-types.h
@@ -267,6 +267,7 @@ struct _ec_lock {
uint32_t refs_pending; /* Refs assigned to fops being prepared */
uint32_t waiting_flags; /*Track xattrop/dirty marking*/
gf_boolean_t acquired;
+ gf_boolean_t contention;
gf_boolean_t unlock_now;
gf_boolean_t release;
gf_boolean_t query;
--
1.8.3.1
|
