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
|
#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
# Copyright (c) 2014-2015, NORDUnet A/S.
# See LICENSE for licensing information.
#
# Distribute the 'sth' file and all missing entries to all frontend nodes.
# See catlfish/doc/merge.txt for more about the merge process.
#
import sys
import json
import requests
import logging
from time import sleep
from base64 import b64encode, b64decode
from os import stat
from multiprocessing import Process, Pipe
from certtools import timing_point, create_ssl_context
from mergetools import get_curpos, get_logorder, chunks, get_missingentries, \
publish_sth, sendlog, sendentries, parse_args, perm, \
get_frontend_verifiedsize, frontend_verify_entries, \
waitforfile, flock_ex_or_fail, Status, loginit, start_worker
def sendlog_helper(entries, curpos, nodename, nodeaddress, own_key, paths,
statusupdates):
logging.info("sending log")
for chunk in chunks(entries, 1000):
for trynumber in range(5, 0, -1):
sendlogresult = sendlog(nodename, nodeaddress,
own_key, paths,
{"start": curpos, "hashes": chunk})
if sendlogresult == None:
if trynumber == 1:
sys.exit(1)
sleep(10)
logging.warning("tries left: %d", trynumber)
continue
break
if sendlogresult["result"] != "ok":
logging.error("sendlog: %s", sendlogresult)
sys.exit(1)
curpos += len(chunk)
statusupdates.status("PROG sending log: %d" % curpos)
logging.info("log sent")
def fill_in_missing_entries(nodename, nodeaddress, own_key, paths, chainsdb,
timing, statusupdates):
missingentries = get_missingentries(nodename, nodeaddress, own_key,
paths)
timing_point(timing, "get missing")
while missingentries:
logging.info("about to send %d missing entries", len(missingentries))
sent_entries = 0
with requests.sessions.Session() as session:
for missingentry_chunk in chunks(missingentries, 100):
missingentry_hashes = [b64decode(missingentry) for missingentry in missingentry_chunk]
hashes_and_entries = [(ehash, chainsdb.get(ehash)) for ehash in missingentry_hashes]
sendentryresult = sendentries(nodename, nodeaddress,
own_key, paths,
hashes_and_entries, session)
if sendentryresult["result"] != "ok":
logging.error("sendentries: %s", sendentryresult)
sys.exit(1)
sent_entries += len(missingentry_hashes)
statusupdates.status(
"PROG sending missing entries: %d" % sent_entries)
timing_point(timing, "send missing")
missingentries = get_missingentries(nodename, nodeaddress,
own_key, paths)
timing_point(timing, "get missing")
def do_send(args, localconfig, frontendnode, logorder, sth, chainsdb, s):
timing = timing_point()
paths = localconfig["paths"]
own_key = (localconfig["nodename"],
"%s/%s-private.pem" % (paths["privatekeys"],
localconfig["nodename"]))
maxwindow = localconfig.get("maxwindow", 1000)
nodename = frontendnode["name"]
nodeaddress = "https://%s/" % frontendnode["address"]
logging.info("distributing for node %s", nodename)
curpos = get_curpos(nodename, nodeaddress, own_key, paths)
timing_point(timing, "get curpos")
logging.info("current position %d", curpos)
verifiedsize = \
get_frontend_verifiedsize(nodename, nodeaddress, own_key, paths)
timing_point(timing, "get verified size")
logging.info("verified size %d", verifiedsize)
assert verifiedsize >= curpos
while verifiedsize < len(logorder):
uptopos = min(verifiedsize + maxwindow, len(logorder))
entries = [b64encode(entry) for entry in logorder[verifiedsize:uptopos]]
sendlog_helper(entries, verifiedsize, nodename, nodeaddress, own_key, paths, s)
timing_point(timing, "sendlog")
fill_in_missing_entries(nodename, nodeaddress, own_key, paths, chainsdb, timing, s)
verifiedsize = frontend_verify_entries(nodename, nodeaddress, own_key, paths, uptopos)
logging.info("sending sth to node %s", nodename)
publishsthresult = publish_sth(nodename, nodeaddress, own_key, paths, sth)
if publishsthresult["result"] != "ok":
logging.info("publishsth: %s", publishsthresult)
sys.exit(1)
timing_point(timing, "send sth")
if args.timing:
logging.debug("timing: merge_dist: %s", timing["deltatimes"])
def merge_dist(args, localconfig, frontendnodes, timestamp):
paths = localconfig["paths"]
mergedb = paths["mergedb"]
chainsdb = perm(localconfig.get("dbbackend", "filedb"), mergedb + "/chains")
logorderfile = mergedb + "/logorder"
sthfile = mergedb + "/sth"
statusfile = mergedb + "/merge_dist.status"
s = Status(statusfile)
create_ssl_context(cafile=paths["https_cacertfile"])
timing = timing_point()
try:
sth = json.loads(open(sthfile, 'r').read())
except (IOError, ValueError):
logging.warning("No valid STH file found in %s", sthfile)
return timestamp, 0
if sth['timestamp'] < timestamp:
logging.warning("New STH file older than the previous one: %d < %d",
sth['timestamp'], timestamp)
return timestamp, 0
if sth['timestamp'] == timestamp:
return timestamp, 0
timestamp = sth['timestamp']
logorder = get_logorder(logorderfile, sth['tree_size'])
timing_point(timing, "get logorder")
procs = {}
for frontendnode in frontendnodes:
nodename = frontendnode["name"]
if args.mergeinterval:
name = 'dist_%s' % nodename
p, pipe = start_worker(name,
lambda _, argv: do_send(*(argv)),
(args, localconfig, frontendnode, logorder, sth, chainsdb, s))
procs[p] = (nodename, pipe)
else:
do_send(args, localconfig, frontendnode, logorder, sth, chainsdb, s)
if not args.mergeinterval:
return timestamp, 0
failures = 0
while True:
for p in list(procs):
if not p.is_alive():
p.join()
nodename, _ = procs[p]
if p.exitcode != 0:
logging.warning("%s failure: %d", nodename, p.exitcode)
failures += 1
del procs[p]
if not procs:
break
sleep(1)
return timestamp, failures
def main():
"""
Wait until 'sth' exists and read it.
Distribute missing entries and the STH to all frontend nodes.
If `--mergeinterval', start over again.
"""
args, config, localconfig = parse_args()
paths = localconfig["paths"]
mergedb = paths["mergedb"]
sth_path = localconfig["paths"]["mergedb"] + "/sth"
lockfile = mergedb + "/.merge_dist.lock"
timestamp = 0
loginit(args, "merge_dist.log")
if not flock_ex_or_fail(lockfile):
logging.critical("unable to take lock %s", lockfile)
return 1
if len(args.node) == 0:
nodes = config["frontendnodes"]
else:
nodes = [n for n in config["frontendnodes"] if n["name"] in args.node]
sth_statinfo = waitforfile(sth_path)
while True:
timestamp, failures = merge_dist(args, localconfig, nodes, timestamp)
if not args.mergeinterval:
break
sth_statinfo_old = sth_statinfo
while sth_statinfo == sth_statinfo_old:
sleep(max(3, args.mergeinterval / 10))
if failures > 0:
break
sth_statinfo = stat(sth_path)
return 0
if __name__ == '__main__':
sys.exit(main())
|
