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
|
// -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
// vim: ts=8 sw=2 smarttab
/*
* Ceph - scalable distributed file system
*
* Copyright (C) 2004-2006 Sage Weil <sage@newdream.net>
*
* This is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License version 2.1, as published by the Free Software
* Foundation. See file COPYING.
*
*/
#ifndef CEPH_MOSDMAP_H
#define CEPH_MOSDMAP_H
#include "msg/Message.h"
#include "osd/OSDMap.h"
#include "include/ceph_features.h"
class MOSDMap final : public Message {
private:
static constexpr int HEAD_VERSION = 4;
static constexpr int COMPAT_VERSION = 3;
public:
uuid_d fsid;
uint64_t encode_features = 0;
std::map<epoch_t, ceph::buffer::list> maps;
std::map<epoch_t, ceph::buffer::list> incremental_maps;
/**
* cluster_osdmap_trim_lower_bound
*
* Encodes a lower bound on the monitor's osdmap trim bound. Recipients
* can safely trim up to this bound. The sender stores maps back to
* cluster_osdmap_trim_lower_bound.
*
* This field was formerly named oldest_map and encoded the oldest map
* stored by the sender. The primary usage of this field, however, was to
* allow the recipient to trim. The secondary usage was to inform the
* recipient of how many maps the sender stored in case it needed to request
* more. For both purposes, it should be safe for an older OSD to interpret
* this field as oldest_map, and it should be safe for a new osd to interpret
* the oldest_map field sent by an older osd as
* cluster_osdmap_trim_lower_bound.
* See bug https://tracker.ceph.com/issues/49689
*/
epoch_t cluster_osdmap_trim_lower_bound = 0;
epoch_t newest_map = 0;
epoch_t get_first() const {
epoch_t e = 0;
auto i = maps.cbegin();
if (i != maps.cend()) e = i->first;
i = incremental_maps.begin();
if (i != incremental_maps.end() &&
(e == 0 || i->first < e)) e = i->first;
return e;
}
epoch_t get_last() const {
epoch_t e = 0;
auto i = maps.crbegin();
if (i != maps.crend()) e = i->first;
i = incremental_maps.rbegin();
if (i != incremental_maps.rend() &&
(e == 0 || i->first > e)) e = i->first;
return e;
}
MOSDMap() : Message{CEPH_MSG_OSD_MAP, HEAD_VERSION, COMPAT_VERSION} { }
MOSDMap(const uuid_d &f, const uint64_t features)
: Message{CEPH_MSG_OSD_MAP, HEAD_VERSION, COMPAT_VERSION},
fsid(f), encode_features(features),
cluster_osdmap_trim_lower_bound(0), newest_map(0) { }
private:
~MOSDMap() final {}
public:
// marshalling
void decode_payload() override {
using ceph::decode;
auto p = payload.cbegin();
decode(fsid, p);
decode(incremental_maps, p);
decode(maps, p);
if (header.version >= 2) {
decode(cluster_osdmap_trim_lower_bound, p);
decode(newest_map, p);
} else {
cluster_osdmap_trim_lower_bound = 0;
newest_map = 0;
}
if (header.version >= 4) {
// removed in octopus
mempool::osdmap::map<int64_t,snap_interval_set_t> gap_removed_snaps;
decode(gap_removed_snaps, p);
}
}
void encode_payload(uint64_t features) override {
using ceph::encode;
header.version = HEAD_VERSION;
header.compat_version = COMPAT_VERSION;
encode(fsid, payload);
if (OSDMap::get_significant_features(encode_features) !=
OSDMap::get_significant_features(features)) {
if ((features & CEPH_FEATURE_PGID64) == 0 ||
(features & CEPH_FEATURE_PGPOOL3) == 0) {
header.version = 1; // old old_client version
header.compat_version = 1;
} else if ((features & CEPH_FEATURE_OSDENC) == 0) {
header.version = 2; // old pg_pool_t
header.compat_version = 2;
}
// reencode maps using old format
//
// FIXME: this can probably be done more efficiently higher up
// the stack, or maybe replaced with something that only
// includes the pools the client cares about.
for (auto p = incremental_maps.begin(); p != incremental_maps.end(); ++p) {
OSDMap::Incremental inc;
auto q = p->second.cbegin();
inc.decode(q);
// always encode with subset of osdmaps canonical features
uint64_t f = inc.encode_features & features;
p->second.clear();
if (inc.fullmap.length()) {
// embedded full std::map?
OSDMap m;
m.decode(inc.fullmap);
inc.fullmap.clear();
m.encode(inc.fullmap, f | CEPH_FEATURE_RESERVED);
}
if (inc.crush.length()) {
// embedded crush std::map
CrushWrapper c;
auto p = inc.crush.cbegin();
c.decode(p);
inc.crush.clear();
c.encode(inc.crush, f);
}
inc.encode(p->second, f | CEPH_FEATURE_RESERVED);
}
for (auto p = maps.begin(); p != maps.end(); ++p) {
OSDMap m;
m.decode(p->second);
// always encode with subset of osdmaps canonical features
uint64_t f = m.get_encoding_features() & features;
p->second.clear();
m.encode(p->second, f | CEPH_FEATURE_RESERVED);
}
}
encode(incremental_maps, payload);
encode(maps, payload);
if (header.version >= 2) {
encode(cluster_osdmap_trim_lower_bound, payload);
encode(newest_map, payload);
}
if (header.version >= 4) {
encode((uint32_t)0, payload);
}
}
std::string_view get_type_name() const override { return "osdmap"; }
void print(std::ostream& out) const override {
out << "osd_map(" << get_first() << ".." << get_last();
if (cluster_osdmap_trim_lower_bound || newest_map)
out << " src has " << cluster_osdmap_trim_lower_bound
<< ".." << newest_map;
out << ")";
}
private:
template<class T, typename... Args>
friend boost::intrusive_ptr<T> ceph::make_message(Args&&... args);
};
#endif
|
