Show More
@@ -1,587 +1,587 b'' | |||||
1 | # fileserverclient.py - client for communicating with the cache process |
|
1 | # fileserverclient.py - client for communicating with the cache process | |
2 | # |
|
2 | # | |
3 | # Copyright 2013 Facebook, Inc. |
|
3 | # Copyright 2013 Facebook, Inc. | |
4 | # |
|
4 | # | |
5 | # This software may be used and distributed according to the terms of the |
|
5 | # This software may be used and distributed according to the terms of the | |
6 | # GNU General Public License version 2 or any later version. |
|
6 | # GNU General Public License version 2 or any later version. | |
7 |
|
7 | |||
8 | from __future__ import absolute_import |
|
8 | from __future__ import absolute_import | |
9 |
|
9 | |||
10 | import hashlib |
|
10 | import hashlib | |
11 | import io |
|
11 | import io | |
12 | import os |
|
12 | import os | |
13 | import threading |
|
13 | import threading | |
14 | import time |
|
14 | import time | |
15 | import zlib |
|
15 | import zlib | |
16 |
|
16 | |||
17 | from mercurial.i18n import _ |
|
17 | from mercurial.i18n import _ | |
18 | from mercurial.node import bin, hex, nullid |
|
18 | from mercurial.node import bin, hex, nullid | |
19 | from mercurial import ( |
|
19 | from mercurial import ( | |
20 | error, |
|
20 | error, | |
21 | node, |
|
21 | node, | |
22 | pycompat, |
|
22 | pycompat, | |
23 | revlog, |
|
23 | revlog, | |
24 | sshpeer, |
|
24 | sshpeer, | |
25 | util, |
|
25 | util, | |
26 | wireprotov1peer, |
|
26 | wireprotov1peer, | |
27 | ) |
|
27 | ) | |
28 | from mercurial.utils import procutil |
|
28 | from mercurial.utils import procutil | |
29 |
|
29 | |||
30 | from . import ( |
|
30 | from . import ( | |
31 | constants, |
|
31 | constants, | |
32 | contentstore, |
|
32 | contentstore, | |
33 | metadatastore, |
|
33 | metadatastore, | |
34 | ) |
|
34 | ) | |
35 |
|
35 | |||
36 | _sshv1peer = sshpeer.sshv1peer |
|
36 | _sshv1peer = sshpeer.sshv1peer | |
37 |
|
37 | |||
38 | # Statistics for debugging |
|
38 | # Statistics for debugging | |
39 | fetchcost = 0 |
|
39 | fetchcost = 0 | |
40 | fetches = 0 |
|
40 | fetches = 0 | |
41 | fetched = 0 |
|
41 | fetched = 0 | |
42 | fetchmisses = 0 |
|
42 | fetchmisses = 0 | |
43 |
|
43 | |||
44 | _lfsmod = None |
|
44 | _lfsmod = None | |
45 |
|
45 | |||
46 | def getcachekey(reponame, file, id): |
|
46 | def getcachekey(reponame, file, id): | |
47 | pathhash = node.hex(hashlib.sha1(file).digest()) |
|
47 | pathhash = node.hex(hashlib.sha1(file).digest()) | |
48 | return os.path.join(reponame, pathhash[:2], pathhash[2:], id) |
|
48 | return os.path.join(reponame, pathhash[:2], pathhash[2:], id) | |
49 |
|
49 | |||
50 | def getlocalkey(file, id): |
|
50 | def getlocalkey(file, id): | |
51 | pathhash = node.hex(hashlib.sha1(file).digest()) |
|
51 | pathhash = node.hex(hashlib.sha1(file).digest()) | |
52 | return os.path.join(pathhash, id) |
|
52 | return os.path.join(pathhash, id) | |
53 |
|
53 | |||
54 | def peersetup(ui, peer): |
|
54 | def peersetup(ui, peer): | |
55 |
|
55 | |||
56 | class remotefilepeer(peer.__class__): |
|
56 | class remotefilepeer(peer.__class__): | |
57 | @wireprotov1peer.batchable |
|
57 | @wireprotov1peer.batchable | |
58 | def x_rfl_getfile(self, file, node): |
|
58 | def x_rfl_getfile(self, file, node): | |
59 | if not self.capable('x_rfl_getfile'): |
|
59 | if not self.capable('x_rfl_getfile'): | |
60 | raise error.Abort( |
|
60 | raise error.Abort( | |
61 | 'configured remotefile server does not support getfile') |
|
61 | 'configured remotefile server does not support getfile') | |
62 | f = wireprotov1peer.future() |
|
62 | f = wireprotov1peer.future() | |
63 | yield {'file': file, 'node': node}, f |
|
63 | yield {'file': file, 'node': node}, f | |
64 | code, data = f.value.split('\0', 1) |
|
64 | code, data = f.value.split('\0', 1) | |
65 | if int(code): |
|
65 | if int(code): | |
66 | raise error.LookupError(file, node, data) |
|
66 | raise error.LookupError(file, node, data) | |
67 | yield data |
|
67 | yield data | |
68 |
|
68 | |||
69 | @wireprotov1peer.batchable |
|
69 | @wireprotov1peer.batchable | |
70 | def x_rfl_getflogheads(self, path): |
|
70 | def x_rfl_getflogheads(self, path): | |
71 | if not self.capable('x_rfl_getflogheads'): |
|
71 | if not self.capable('x_rfl_getflogheads'): | |
72 | raise error.Abort('configured remotefile server does not ' |
|
72 | raise error.Abort('configured remotefile server does not ' | |
73 | 'support getflogheads') |
|
73 | 'support getflogheads') | |
74 | f = wireprotov1peer.future() |
|
74 | f = wireprotov1peer.future() | |
75 | yield {'path': path}, f |
|
75 | yield {'path': path}, f | |
76 | heads = f.value.split('\n') if f.value else [] |
|
76 | heads = f.value.split('\n') if f.value else [] | |
77 | yield heads |
|
77 | yield heads | |
78 |
|
78 | |||
79 | def _updatecallstreamopts(self, command, opts): |
|
79 | def _updatecallstreamopts(self, command, opts): | |
80 | if command != 'getbundle': |
|
80 | if command != 'getbundle': | |
81 | return |
|
81 | return | |
82 | if (constants.NETWORK_CAP_LEGACY_SSH_GETFILES |
|
82 | if (constants.NETWORK_CAP_LEGACY_SSH_GETFILES | |
83 | not in self.capabilities()): |
|
83 | not in self.capabilities()): | |
84 | return |
|
84 | return | |
85 | if not util.safehasattr(self, '_localrepo'): |
|
85 | if not util.safehasattr(self, '_localrepo'): | |
86 | return |
|
86 | return | |
87 | if (constants.SHALLOWREPO_REQUIREMENT |
|
87 | if (constants.SHALLOWREPO_REQUIREMENT | |
88 | not in self._localrepo.requirements): |
|
88 | not in self._localrepo.requirements): | |
89 | return |
|
89 | return | |
90 |
|
90 | |||
91 | bundlecaps = opts.get('bundlecaps') |
|
91 | bundlecaps = opts.get('bundlecaps') | |
92 | if bundlecaps: |
|
92 | if bundlecaps: | |
93 | bundlecaps = [bundlecaps] |
|
93 | bundlecaps = [bundlecaps] | |
94 | else: |
|
94 | else: | |
95 | bundlecaps = [] |
|
95 | bundlecaps = [] | |
96 |
|
96 | |||
97 | # shallow, includepattern, and excludepattern are a hacky way of |
|
97 | # shallow, includepattern, and excludepattern are a hacky way of | |
98 | # carrying over data from the local repo to this getbundle |
|
98 | # carrying over data from the local repo to this getbundle | |
99 | # command. We need to do it this way because bundle1 getbundle |
|
99 | # command. We need to do it this way because bundle1 getbundle | |
100 | # doesn't provide any other place we can hook in to manipulate |
|
100 | # doesn't provide any other place we can hook in to manipulate | |
101 | # getbundle args before it goes across the wire. Once we get rid |
|
101 | # getbundle args before it goes across the wire. Once we get rid | |
102 | # of bundle1, we can use bundle2's _pullbundle2extraprepare to |
|
102 | # of bundle1, we can use bundle2's _pullbundle2extraprepare to | |
103 | # do this more cleanly. |
|
103 | # do this more cleanly. | |
104 | bundlecaps.append(constants.BUNDLE2_CAPABLITY) |
|
104 | bundlecaps.append(constants.BUNDLE2_CAPABLITY) | |
105 | if self._localrepo.includepattern: |
|
105 | if self._localrepo.includepattern: | |
106 | patterns = '\0'.join(self._localrepo.includepattern) |
|
106 | patterns = '\0'.join(self._localrepo.includepattern) | |
107 | includecap = "includepattern=" + patterns |
|
107 | includecap = "includepattern=" + patterns | |
108 | bundlecaps.append(includecap) |
|
108 | bundlecaps.append(includecap) | |
109 | if self._localrepo.excludepattern: |
|
109 | if self._localrepo.excludepattern: | |
110 | patterns = '\0'.join(self._localrepo.excludepattern) |
|
110 | patterns = '\0'.join(self._localrepo.excludepattern) | |
111 | excludecap = "excludepattern=" + patterns |
|
111 | excludecap = "excludepattern=" + patterns | |
112 | bundlecaps.append(excludecap) |
|
112 | bundlecaps.append(excludecap) | |
113 | opts['bundlecaps'] = ','.join(bundlecaps) |
|
113 | opts['bundlecaps'] = ','.join(bundlecaps) | |
114 |
|
114 | |||
115 | def _sendrequest(self, command, args, **opts): |
|
115 | def _sendrequest(self, command, args, **opts): | |
116 | self._updatecallstreamopts(command, args) |
|
116 | self._updatecallstreamopts(command, args) | |
117 | return super(remotefilepeer, self)._sendrequest(command, args, |
|
117 | return super(remotefilepeer, self)._sendrequest(command, args, | |
118 | **opts) |
|
118 | **opts) | |
119 |
|
119 | |||
120 | def _callstream(self, command, **opts): |
|
120 | def _callstream(self, command, **opts): | |
121 | supertype = super(remotefilepeer, self) |
|
121 | supertype = super(remotefilepeer, self) | |
122 | if not util.safehasattr(supertype, '_sendrequest'): |
|
122 | if not util.safehasattr(supertype, '_sendrequest'): | |
123 | self._updatecallstreamopts(command, pycompat.byteskwargs(opts)) |
|
123 | self._updatecallstreamopts(command, pycompat.byteskwargs(opts)) | |
124 | return super(remotefilepeer, self)._callstream(command, **opts) |
|
124 | return super(remotefilepeer, self)._callstream(command, **opts) | |
125 |
|
125 | |||
126 | peer.__class__ = remotefilepeer |
|
126 | peer.__class__ = remotefilepeer | |
127 |
|
127 | |||
128 | class cacheconnection(object): |
|
128 | class cacheconnection(object): | |
129 | """The connection for communicating with the remote cache. Performs |
|
129 | """The connection for communicating with the remote cache. Performs | |
130 | gets and sets by communicating with an external process that has the |
|
130 | gets and sets by communicating with an external process that has the | |
131 | cache-specific implementation. |
|
131 | cache-specific implementation. | |
132 | """ |
|
132 | """ | |
133 | def __init__(self): |
|
133 | def __init__(self): | |
134 | self.pipeo = self.pipei = self.pipee = None |
|
134 | self.pipeo = self.pipei = self.pipee = None | |
135 | self.subprocess = None |
|
135 | self.subprocess = None | |
136 | self.connected = False |
|
136 | self.connected = False | |
137 |
|
137 | |||
138 | def connect(self, cachecommand): |
|
138 | def connect(self, cachecommand): | |
139 | if self.pipeo: |
|
139 | if self.pipeo: | |
140 | raise error.Abort(_("cache connection already open")) |
|
140 | raise error.Abort(_("cache connection already open")) | |
141 | self.pipei, self.pipeo, self.pipee, self.subprocess = \ |
|
141 | self.pipei, self.pipeo, self.pipee, self.subprocess = \ | |
142 | procutil.popen4(cachecommand) |
|
142 | procutil.popen4(cachecommand) | |
143 | self.connected = True |
|
143 | self.connected = True | |
144 |
|
144 | |||
145 | def close(self): |
|
145 | def close(self): | |
146 | def tryclose(pipe): |
|
146 | def tryclose(pipe): | |
147 | try: |
|
147 | try: | |
148 | pipe.close() |
|
148 | pipe.close() | |
149 | except Exception: |
|
149 | except Exception: | |
150 | pass |
|
150 | pass | |
151 | if self.connected: |
|
151 | if self.connected: | |
152 | try: |
|
152 | try: | |
153 | self.pipei.write("exit\n") |
|
153 | self.pipei.write("exit\n") | |
154 | except Exception: |
|
154 | except Exception: | |
155 | pass |
|
155 | pass | |
156 | tryclose(self.pipei) |
|
156 | tryclose(self.pipei) | |
157 | self.pipei = None |
|
157 | self.pipei = None | |
158 | tryclose(self.pipeo) |
|
158 | tryclose(self.pipeo) | |
159 | self.pipeo = None |
|
159 | self.pipeo = None | |
160 | tryclose(self.pipee) |
|
160 | tryclose(self.pipee) | |
161 | self.pipee = None |
|
161 | self.pipee = None | |
162 | try: |
|
162 | try: | |
163 | # Wait for process to terminate, making sure to avoid deadlock. |
|
163 | # Wait for process to terminate, making sure to avoid deadlock. | |
164 | # See https://docs.python.org/2/library/subprocess.html for |
|
164 | # See https://docs.python.org/2/library/subprocess.html for | |
165 | # warnings about wait() and deadlocking. |
|
165 | # warnings about wait() and deadlocking. | |
166 | self.subprocess.communicate() |
|
166 | self.subprocess.communicate() | |
167 | except Exception: |
|
167 | except Exception: | |
168 | pass |
|
168 | pass | |
169 | self.subprocess = None |
|
169 | self.subprocess = None | |
170 | self.connected = False |
|
170 | self.connected = False | |
171 |
|
171 | |||
172 | def request(self, request, flush=True): |
|
172 | def request(self, request, flush=True): | |
173 | if self.connected: |
|
173 | if self.connected: | |
174 | try: |
|
174 | try: | |
175 | self.pipei.write(request) |
|
175 | self.pipei.write(request) | |
176 | if flush: |
|
176 | if flush: | |
177 | self.pipei.flush() |
|
177 | self.pipei.flush() | |
178 | except IOError: |
|
178 | except IOError: | |
179 | self.close() |
|
179 | self.close() | |
180 |
|
180 | |||
181 | def receiveline(self): |
|
181 | def receiveline(self): | |
182 | if not self.connected: |
|
182 | if not self.connected: | |
183 | return None |
|
183 | return None | |
184 | try: |
|
184 | try: | |
185 | result = self.pipeo.readline()[:-1] |
|
185 | result = self.pipeo.readline()[:-1] | |
186 | if not result: |
|
186 | if not result: | |
187 | self.close() |
|
187 | self.close() | |
188 | except IOError: |
|
188 | except IOError: | |
189 | self.close() |
|
189 | self.close() | |
190 |
|
190 | |||
191 | return result |
|
191 | return result | |
192 |
|
192 | |||
193 | def _getfilesbatch( |
|
193 | def _getfilesbatch( | |
194 | remote, receivemissing, progresstick, missed, idmap, batchsize): |
|
194 | remote, receivemissing, progresstick, missed, idmap, batchsize): | |
195 | # Over http(s), iterbatch is a streamy method and we can start |
|
195 | # Over http(s), iterbatch is a streamy method and we can start | |
196 | # looking at results early. This means we send one (potentially |
|
196 | # looking at results early. This means we send one (potentially | |
197 | # large) request, but then we show nice progress as we process |
|
197 | # large) request, but then we show nice progress as we process | |
198 | # file results, rather than showing chunks of $batchsize in |
|
198 | # file results, rather than showing chunks of $batchsize in | |
199 | # progress. |
|
199 | # progress. | |
200 | # |
|
200 | # | |
201 | # Over ssh, iterbatch isn't streamy because batch() wasn't |
|
201 | # Over ssh, iterbatch isn't streamy because batch() wasn't | |
202 | # explicitly designed as a streaming method. In the future we |
|
202 | # explicitly designed as a streaming method. In the future we | |
203 | # should probably introduce a streambatch() method upstream and |
|
203 | # should probably introduce a streambatch() method upstream and | |
204 | # use that for this. |
|
204 | # use that for this. | |
205 | with remote.commandexecutor() as e: |
|
205 | with remote.commandexecutor() as e: | |
206 | futures = [] |
|
206 | futures = [] | |
207 | for m in missed: |
|
207 | for m in missed: | |
208 | futures.append(e.callcommand('x_rfl_getfile', { |
|
208 | futures.append(e.callcommand('x_rfl_getfile', { | |
209 | 'file': idmap[m], |
|
209 | 'file': idmap[m], | |
210 | 'node': m[-40:] |
|
210 | 'node': m[-40:] | |
211 | })) |
|
211 | })) | |
212 |
|
212 | |||
213 | for i, m in enumerate(missed): |
|
213 | for i, m in enumerate(missed): | |
214 | r = futures[i].result() |
|
214 | r = futures[i].result() | |
215 | futures[i] = None # release memory |
|
215 | futures[i] = None # release memory | |
216 | file_ = idmap[m] |
|
216 | file_ = idmap[m] | |
217 | node = m[-40:] |
|
217 | node = m[-40:] | |
218 | receivemissing(io.BytesIO('%d\n%s' % (len(r), r)), file_, node) |
|
218 | receivemissing(io.BytesIO('%d\n%s' % (len(r), r)), file_, node) | |
219 | progresstick() |
|
219 | progresstick() | |
220 |
|
220 | |||
221 | def _getfiles_optimistic( |
|
221 | def _getfiles_optimistic( | |
222 | remote, receivemissing, progresstick, missed, idmap, step): |
|
222 | remote, receivemissing, progresstick, missed, idmap, step): | |
223 | remote._callstream("x_rfl_getfiles") |
|
223 | remote._callstream("x_rfl_getfiles") | |
224 | i = 0 |
|
224 | i = 0 | |
225 | pipeo = remote._pipeo |
|
225 | pipeo = remote._pipeo | |
226 | pipei = remote._pipei |
|
226 | pipei = remote._pipei | |
227 | while i < len(missed): |
|
227 | while i < len(missed): | |
228 | # issue a batch of requests |
|
228 | # issue a batch of requests | |
229 | start = i |
|
229 | start = i | |
230 | end = min(len(missed), start + step) |
|
230 | end = min(len(missed), start + step) | |
231 | i = end |
|
231 | i = end | |
232 | for missingid in missed[start:end]: |
|
232 | for missingid in missed[start:end]: | |
233 | # issue new request |
|
233 | # issue new request | |
234 | versionid = missingid[-40:] |
|
234 | versionid = missingid[-40:] | |
235 | file = idmap[missingid] |
|
235 | file = idmap[missingid] | |
236 | sshrequest = "%s%s\n" % (versionid, file) |
|
236 | sshrequest = "%s%s\n" % (versionid, file) | |
237 | pipeo.write(sshrequest) |
|
237 | pipeo.write(sshrequest) | |
238 | pipeo.flush() |
|
238 | pipeo.flush() | |
239 |
|
239 | |||
240 | # receive batch results |
|
240 | # receive batch results | |
241 | for missingid in missed[start:end]: |
|
241 | for missingid in missed[start:end]: | |
242 | versionid = missingid[-40:] |
|
242 | versionid = missingid[-40:] | |
243 | file = idmap[missingid] |
|
243 | file = idmap[missingid] | |
244 | receivemissing(pipei, file, versionid) |
|
244 | receivemissing(pipei, file, versionid) | |
245 | progresstick() |
|
245 | progresstick() | |
246 |
|
246 | |||
247 | # End the command |
|
247 | # End the command | |
248 | pipeo.write('\n') |
|
248 | pipeo.write('\n') | |
249 | pipeo.flush() |
|
249 | pipeo.flush() | |
250 |
|
250 | |||
251 | def _getfiles_threaded( |
|
251 | def _getfiles_threaded( | |
252 | remote, receivemissing, progresstick, missed, idmap, step): |
|
252 | remote, receivemissing, progresstick, missed, idmap, step): | |
253 | remote._callstream("getfiles") |
|
253 | remote._callstream("getfiles") | |
254 | pipeo = remote._pipeo |
|
254 | pipeo = remote._pipeo | |
255 | pipei = remote._pipei |
|
255 | pipei = remote._pipei | |
256 |
|
256 | |||
257 | def writer(): |
|
257 | def writer(): | |
258 | for missingid in missed: |
|
258 | for missingid in missed: | |
259 | versionid = missingid[-40:] |
|
259 | versionid = missingid[-40:] | |
260 | file = idmap[missingid] |
|
260 | file = idmap[missingid] | |
261 | sshrequest = "%s%s\n" % (versionid, file) |
|
261 | sshrequest = "%s%s\n" % (versionid, file) | |
262 | pipeo.write(sshrequest) |
|
262 | pipeo.write(sshrequest) | |
263 | pipeo.flush() |
|
263 | pipeo.flush() | |
264 | writerthread = threading.Thread(target=writer) |
|
264 | writerthread = threading.Thread(target=writer) | |
265 | writerthread.daemon = True |
|
265 | writerthread.daemon = True | |
266 | writerthread.start() |
|
266 | writerthread.start() | |
267 |
|
267 | |||
268 | for missingid in missed: |
|
268 | for missingid in missed: | |
269 | versionid = missingid[-40:] |
|
269 | versionid = missingid[-40:] | |
270 | file = idmap[missingid] |
|
270 | file = idmap[missingid] | |
271 | receivemissing(pipei, file, versionid) |
|
271 | receivemissing(pipei, file, versionid) | |
272 | progresstick() |
|
272 | progresstick() | |
273 |
|
273 | |||
274 | writerthread.join() |
|
274 | writerthread.join() | |
275 | # End the command |
|
275 | # End the command | |
276 | pipeo.write('\n') |
|
276 | pipeo.write('\n') | |
277 | pipeo.flush() |
|
277 | pipeo.flush() | |
278 |
|
278 | |||
279 | class fileserverclient(object): |
|
279 | class fileserverclient(object): | |
280 | """A client for requesting files from the remote file server. |
|
280 | """A client for requesting files from the remote file server. | |
281 | """ |
|
281 | """ | |
282 | def __init__(self, repo): |
|
282 | def __init__(self, repo): | |
283 | ui = repo.ui |
|
283 | ui = repo.ui | |
284 | self.repo = repo |
|
284 | self.repo = repo | |
285 | self.ui = ui |
|
285 | self.ui = ui | |
286 | self.cacheprocess = ui.config("remotefilelog", "cacheprocess") |
|
286 | self.cacheprocess = ui.config("remotefilelog", "cacheprocess") | |
287 | if self.cacheprocess: |
|
287 | if self.cacheprocess: | |
288 | self.cacheprocess = util.expandpath(self.cacheprocess) |
|
288 | self.cacheprocess = util.expandpath(self.cacheprocess) | |
289 |
|
289 | |||
290 | # This option causes remotefilelog to pass the full file path to the |
|
290 | # This option causes remotefilelog to pass the full file path to the | |
291 | # cacheprocess instead of a hashed key. |
|
291 | # cacheprocess instead of a hashed key. | |
292 | self.cacheprocesspasspath = ui.configbool( |
|
292 | self.cacheprocesspasspath = ui.configbool( | |
293 | "remotefilelog", "cacheprocess.includepath") |
|
293 | "remotefilelog", "cacheprocess.includepath") | |
294 |
|
294 | |||
295 | self.debugoutput = ui.configbool("remotefilelog", "debug") |
|
295 | self.debugoutput = ui.configbool("remotefilelog", "debug") | |
296 |
|
296 | |||
297 | self.remotecache = cacheconnection() |
|
297 | self.remotecache = cacheconnection() | |
298 |
|
298 | |||
299 | def setstore(self, datastore, historystore, writedata, writehistory): |
|
299 | def setstore(self, datastore, historystore, writedata, writehistory): | |
300 | self.datastore = datastore |
|
300 | self.datastore = datastore | |
301 | self.historystore = historystore |
|
301 | self.historystore = historystore | |
302 | self.writedata = writedata |
|
302 | self.writedata = writedata | |
303 | self.writehistory = writehistory |
|
303 | self.writehistory = writehistory | |
304 |
|
304 | |||
305 | def _connect(self): |
|
305 | def _connect(self): | |
306 | return self.repo.connectionpool.get(self.repo.fallbackpath) |
|
306 | return self.repo.connectionpool.get(self.repo.fallbackpath) | |
307 |
|
307 | |||
308 | def request(self, fileids): |
|
308 | def request(self, fileids): | |
309 | """Takes a list of filename/node pairs and fetches them from the |
|
309 | """Takes a list of filename/node pairs and fetches them from the | |
310 | server. Files are stored in the local cache. |
|
310 | server. Files are stored in the local cache. | |
311 | A list of nodes that the server couldn't find is returned. |
|
311 | A list of nodes that the server couldn't find is returned. | |
312 | If the connection fails, an exception is raised. |
|
312 | If the connection fails, an exception is raised. | |
313 | """ |
|
313 | """ | |
314 | if not self.remotecache.connected: |
|
314 | if not self.remotecache.connected: | |
315 | self.connect() |
|
315 | self.connect() | |
316 | cache = self.remotecache |
|
316 | cache = self.remotecache | |
317 | writedata = self.writedata |
|
317 | writedata = self.writedata | |
318 |
|
318 | |||
319 | repo = self.repo |
|
319 | repo = self.repo | |
320 | total = len(fileids) |
|
320 | total = len(fileids) | |
321 | request = "get\n%d\n" % total |
|
321 | request = "get\n%d\n" % total | |
322 | idmap = {} |
|
322 | idmap = {} | |
323 | reponame = repo.name |
|
323 | reponame = repo.name | |
324 | for file, id in fileids: |
|
324 | for file, id in fileids: | |
325 | fullid = getcachekey(reponame, file, id) |
|
325 | fullid = getcachekey(reponame, file, id) | |
326 | if self.cacheprocesspasspath: |
|
326 | if self.cacheprocesspasspath: | |
327 | request += file + '\0' |
|
327 | request += file + '\0' | |
328 | request += fullid + "\n" |
|
328 | request += fullid + "\n" | |
329 | idmap[fullid] = file |
|
329 | idmap[fullid] = file | |
330 |
|
330 | |||
331 | cache.request(request) |
|
331 | cache.request(request) | |
332 |
|
332 | |||
333 | progress = self.ui.makeprogress(_('downloading'), total=total) |
|
333 | progress = self.ui.makeprogress(_('downloading'), total=total) | |
334 | progress.update(0) |
|
334 | progress.update(0) | |
335 |
|
335 | |||
336 | missed = [] |
|
336 | missed = [] | |
337 | count = 0 |
|
337 | count = 0 | |
338 | while True: |
|
338 | while True: | |
339 | missingid = cache.receiveline() |
|
339 | missingid = cache.receiveline() | |
340 | if not missingid: |
|
340 | if not missingid: | |
341 | missedset = set(missed) |
|
341 | missedset = set(missed) | |
342 | for missingid in idmap: |
|
342 | for missingid in idmap: | |
343 | if not missingid in missedset: |
|
343 | if not missingid in missedset: | |
344 | missed.append(missingid) |
|
344 | missed.append(missingid) | |
345 | self.ui.warn(_("warning: cache connection closed early - " + |
|
345 | self.ui.warn(_("warning: cache connection closed early - " + | |
346 | "falling back to server\n")) |
|
346 | "falling back to server\n")) | |
347 | break |
|
347 | break | |
348 | if missingid == "0": |
|
348 | if missingid == "0": | |
349 | break |
|
349 | break | |
350 | if missingid.startswith("_hits_"): |
|
350 | if missingid.startswith("_hits_"): | |
351 | # receive progress reports |
|
351 | # receive progress reports | |
352 | parts = missingid.split("_") |
|
352 | parts = missingid.split("_") | |
353 | count += int(parts[2]) |
|
353 | count += int(parts[2]) | |
354 | progress.update(count) |
|
354 | progress.update(count) | |
355 | continue |
|
355 | continue | |
356 |
|
356 | |||
357 | missed.append(missingid) |
|
357 | missed.append(missingid) | |
358 |
|
358 | |||
359 | global fetchmisses |
|
359 | global fetchmisses | |
360 | fetchmisses += len(missed) |
|
360 | fetchmisses += len(missed) | |
361 |
|
361 | |||
362 |
|
|
362 | fromcache = total - len(missed) | |
363 |
|
|
363 | count = [fromcache] | |
364 | progress.update(count[0], total=total) |
|
364 | progress.update(count[0], total=total) | |
365 | self.ui.log("remotefilelog", "remote cache hit rate is %r of %r\n", |
|
365 | self.ui.log("remotefilelog", "remote cache hit rate is %r of %r\n", | |
366 | count[0], total, hit=count[0], total=total) |
|
366 | count[0], total, hit=count[0], total=total) | |
367 |
|
367 | |||
368 | oldumask = os.umask(0o002) |
|
368 | oldumask = os.umask(0o002) | |
369 | try: |
|
369 | try: | |
370 | # receive cache misses from master |
|
370 | # receive cache misses from master | |
371 | if missed: |
|
371 | if missed: | |
372 | def progresstick(): |
|
372 | def progresstick(): | |
373 | count[0] += 1 |
|
373 | count[0] += 1 | |
374 | progress.update(count[0]) |
|
374 | progress.update(count[0]) | |
375 | # When verbose is true, sshpeer prints 'running ssh...' |
|
375 | # When verbose is true, sshpeer prints 'running ssh...' | |
376 | # to stdout, which can interfere with some command |
|
376 | # to stdout, which can interfere with some command | |
377 | # outputs |
|
377 | # outputs | |
378 | verbose = self.ui.verbose |
|
378 | verbose = self.ui.verbose | |
379 | self.ui.verbose = False |
|
379 | self.ui.verbose = False | |
380 | try: |
|
380 | try: | |
381 | with self._connect() as conn: |
|
381 | with self._connect() as conn: | |
382 | remote = conn.peer |
|
382 | remote = conn.peer | |
383 | if remote.capable( |
|
383 | if remote.capable( | |
384 | constants.NETWORK_CAP_LEGACY_SSH_GETFILES): |
|
384 | constants.NETWORK_CAP_LEGACY_SSH_GETFILES): | |
385 | if not isinstance(remote, _sshv1peer): |
|
385 | if not isinstance(remote, _sshv1peer): | |
386 | raise error.Abort('remotefilelog requires ssh ' |
|
386 | raise error.Abort('remotefilelog requires ssh ' | |
387 | 'servers') |
|
387 | 'servers') | |
388 | step = self.ui.configint('remotefilelog', |
|
388 | step = self.ui.configint('remotefilelog', | |
389 | 'getfilesstep') |
|
389 | 'getfilesstep') | |
390 | getfilestype = self.ui.config('remotefilelog', |
|
390 | getfilestype = self.ui.config('remotefilelog', | |
391 | 'getfilestype') |
|
391 | 'getfilestype') | |
392 | if getfilestype == 'threaded': |
|
392 | if getfilestype == 'threaded': | |
393 | _getfiles = _getfiles_threaded |
|
393 | _getfiles = _getfiles_threaded | |
394 | else: |
|
394 | else: | |
395 | _getfiles = _getfiles_optimistic |
|
395 | _getfiles = _getfiles_optimistic | |
396 | _getfiles(remote, self.receivemissing, progresstick, |
|
396 | _getfiles(remote, self.receivemissing, progresstick, | |
397 | missed, idmap, step) |
|
397 | missed, idmap, step) | |
398 | elif remote.capable("x_rfl_getfile"): |
|
398 | elif remote.capable("x_rfl_getfile"): | |
399 | if remote.capable('batch'): |
|
399 | if remote.capable('batch'): | |
400 | batchdefault = 100 |
|
400 | batchdefault = 100 | |
401 | else: |
|
401 | else: | |
402 | batchdefault = 10 |
|
402 | batchdefault = 10 | |
403 | batchsize = self.ui.configint( |
|
403 | batchsize = self.ui.configint( | |
404 | 'remotefilelog', 'batchsize', batchdefault) |
|
404 | 'remotefilelog', 'batchsize', batchdefault) | |
405 | _getfilesbatch( |
|
405 | _getfilesbatch( | |
406 | remote, self.receivemissing, progresstick, |
|
406 | remote, self.receivemissing, progresstick, | |
407 | missed, idmap, batchsize) |
|
407 | missed, idmap, batchsize) | |
408 | else: |
|
408 | else: | |
409 | raise error.Abort("configured remotefilelog server" |
|
409 | raise error.Abort("configured remotefilelog server" | |
410 | " does not support remotefilelog") |
|
410 | " does not support remotefilelog") | |
411 |
|
411 | |||
412 | self.ui.log("remotefilefetchlog", |
|
412 | self.ui.log("remotefilefetchlog", | |
413 | "Success\n", |
|
413 | "Success\n", | |
414 | fetched_files = count[0] - fromcache, |
|
414 | fetched_files = count[0] - fromcache, | |
415 | total_to_fetch = total - fromcache) |
|
415 | total_to_fetch = total - fromcache) | |
416 | except Exception: |
|
416 | except Exception: | |
417 | self.ui.log("remotefilefetchlog", |
|
417 | self.ui.log("remotefilefetchlog", | |
418 | "Fail\n", |
|
418 | "Fail\n", | |
419 | fetched_files = count[0] - fromcache, |
|
419 | fetched_files = count[0] - fromcache, | |
420 | total_to_fetch = total - fromcache) |
|
420 | total_to_fetch = total - fromcache) | |
421 | raise |
|
421 | raise | |
422 | finally: |
|
422 | finally: | |
423 | self.ui.verbose = verbose |
|
423 | self.ui.verbose = verbose | |
424 | # send to memcache |
|
424 | # send to memcache | |
425 | request = "set\n%d\n%s\n" % (len(missed), "\n".join(missed)) |
|
425 | request = "set\n%d\n%s\n" % (len(missed), "\n".join(missed)) | |
426 | cache.request(request) |
|
426 | cache.request(request) | |
427 |
|
427 | |||
428 | progress.complete() |
|
428 | progress.complete() | |
429 |
|
429 | |||
430 | # mark ourselves as a user of this cache |
|
430 | # mark ourselves as a user of this cache | |
431 | writedata.markrepo(self.repo.path) |
|
431 | writedata.markrepo(self.repo.path) | |
432 | finally: |
|
432 | finally: | |
433 | os.umask(oldumask) |
|
433 | os.umask(oldumask) | |
434 |
|
434 | |||
435 | def receivemissing(self, pipe, filename, node): |
|
435 | def receivemissing(self, pipe, filename, node): | |
436 | line = pipe.readline()[:-1] |
|
436 | line = pipe.readline()[:-1] | |
437 | if not line: |
|
437 | if not line: | |
438 | raise error.ResponseError(_("error downloading file contents:"), |
|
438 | raise error.ResponseError(_("error downloading file contents:"), | |
439 | _("connection closed early")) |
|
439 | _("connection closed early")) | |
440 | size = int(line) |
|
440 | size = int(line) | |
441 | data = pipe.read(size) |
|
441 | data = pipe.read(size) | |
442 | if len(data) != size: |
|
442 | if len(data) != size: | |
443 | raise error.ResponseError(_("error downloading file contents:"), |
|
443 | raise error.ResponseError(_("error downloading file contents:"), | |
444 | _("only received %s of %s bytes") |
|
444 | _("only received %s of %s bytes") | |
445 | % (len(data), size)) |
|
445 | % (len(data), size)) | |
446 |
|
446 | |||
447 | self.writedata.addremotefilelognode(filename, bin(node), |
|
447 | self.writedata.addremotefilelognode(filename, bin(node), | |
448 | zlib.decompress(data)) |
|
448 | zlib.decompress(data)) | |
449 |
|
449 | |||
450 | def connect(self): |
|
450 | def connect(self): | |
451 | if self.cacheprocess: |
|
451 | if self.cacheprocess: | |
452 | cmd = "%s %s" % (self.cacheprocess, self.writedata._path) |
|
452 | cmd = "%s %s" % (self.cacheprocess, self.writedata._path) | |
453 | self.remotecache.connect(cmd) |
|
453 | self.remotecache.connect(cmd) | |
454 | else: |
|
454 | else: | |
455 | # If no cache process is specified, we fake one that always |
|
455 | # If no cache process is specified, we fake one that always | |
456 | # returns cache misses. This enables tests to run easily |
|
456 | # returns cache misses. This enables tests to run easily | |
457 | # and may eventually allow us to be a drop in replacement |
|
457 | # and may eventually allow us to be a drop in replacement | |
458 | # for the largefiles extension. |
|
458 | # for the largefiles extension. | |
459 | class simplecache(object): |
|
459 | class simplecache(object): | |
460 | def __init__(self): |
|
460 | def __init__(self): | |
461 | self.missingids = [] |
|
461 | self.missingids = [] | |
462 | self.connected = True |
|
462 | self.connected = True | |
463 |
|
463 | |||
464 | def close(self): |
|
464 | def close(self): | |
465 | pass |
|
465 | pass | |
466 |
|
466 | |||
467 | def request(self, value, flush=True): |
|
467 | def request(self, value, flush=True): | |
468 | lines = value.split("\n") |
|
468 | lines = value.split("\n") | |
469 | if lines[0] != "get": |
|
469 | if lines[0] != "get": | |
470 | return |
|
470 | return | |
471 | self.missingids = lines[2:-1] |
|
471 | self.missingids = lines[2:-1] | |
472 | self.missingids.append('0') |
|
472 | self.missingids.append('0') | |
473 |
|
473 | |||
474 | def receiveline(self): |
|
474 | def receiveline(self): | |
475 | if len(self.missingids) > 0: |
|
475 | if len(self.missingids) > 0: | |
476 | return self.missingids.pop(0) |
|
476 | return self.missingids.pop(0) | |
477 | return None |
|
477 | return None | |
478 |
|
478 | |||
479 | self.remotecache = simplecache() |
|
479 | self.remotecache = simplecache() | |
480 |
|
480 | |||
481 | def close(self): |
|
481 | def close(self): | |
482 | if fetches: |
|
482 | if fetches: | |
483 | msg = ("%d files fetched over %d fetches - " + |
|
483 | msg = ("%d files fetched over %d fetches - " + | |
484 | "(%d misses, %0.2f%% hit ratio) over %0.2fs\n") % ( |
|
484 | "(%d misses, %0.2f%% hit ratio) over %0.2fs\n") % ( | |
485 | fetched, |
|
485 | fetched, | |
486 | fetches, |
|
486 | fetches, | |
487 | fetchmisses, |
|
487 | fetchmisses, | |
488 | float(fetched - fetchmisses) / float(fetched) * 100.0, |
|
488 | float(fetched - fetchmisses) / float(fetched) * 100.0, | |
489 | fetchcost) |
|
489 | fetchcost) | |
490 | if self.debugoutput: |
|
490 | if self.debugoutput: | |
491 | self.ui.warn(msg) |
|
491 | self.ui.warn(msg) | |
492 | self.ui.log("remotefilelog.prefetch", msg.replace("%", "%%"), |
|
492 | self.ui.log("remotefilelog.prefetch", msg.replace("%", "%%"), | |
493 | remotefilelogfetched=fetched, |
|
493 | remotefilelogfetched=fetched, | |
494 | remotefilelogfetches=fetches, |
|
494 | remotefilelogfetches=fetches, | |
495 | remotefilelogfetchmisses=fetchmisses, |
|
495 | remotefilelogfetchmisses=fetchmisses, | |
496 | remotefilelogfetchtime=fetchcost * 1000) |
|
496 | remotefilelogfetchtime=fetchcost * 1000) | |
497 |
|
497 | |||
498 | if self.remotecache.connected: |
|
498 | if self.remotecache.connected: | |
499 | self.remotecache.close() |
|
499 | self.remotecache.close() | |
500 |
|
500 | |||
501 | def prefetch(self, fileids, force=False, fetchdata=True, |
|
501 | def prefetch(self, fileids, force=False, fetchdata=True, | |
502 | fetchhistory=False): |
|
502 | fetchhistory=False): | |
503 | """downloads the given file versions to the cache |
|
503 | """downloads the given file versions to the cache | |
504 | """ |
|
504 | """ | |
505 | repo = self.repo |
|
505 | repo = self.repo | |
506 | idstocheck = [] |
|
506 | idstocheck = [] | |
507 | for file, id in fileids: |
|
507 | for file, id in fileids: | |
508 | # hack |
|
508 | # hack | |
509 | # - we don't use .hgtags |
|
509 | # - we don't use .hgtags | |
510 | # - workingctx produces ids with length 42, |
|
510 | # - workingctx produces ids with length 42, | |
511 | # which we skip since they aren't in any cache |
|
511 | # which we skip since they aren't in any cache | |
512 | if (file == '.hgtags' or len(id) == 42 |
|
512 | if (file == '.hgtags' or len(id) == 42 | |
513 | or not repo.shallowmatch(file)): |
|
513 | or not repo.shallowmatch(file)): | |
514 | continue |
|
514 | continue | |
515 |
|
515 | |||
516 | idstocheck.append((file, bin(id))) |
|
516 | idstocheck.append((file, bin(id))) | |
517 |
|
517 | |||
518 | datastore = self.datastore |
|
518 | datastore = self.datastore | |
519 | historystore = self.historystore |
|
519 | historystore = self.historystore | |
520 | if force: |
|
520 | if force: | |
521 | datastore = contentstore.unioncontentstore(*repo.shareddatastores) |
|
521 | datastore = contentstore.unioncontentstore(*repo.shareddatastores) | |
522 | historystore = metadatastore.unionmetadatastore( |
|
522 | historystore = metadatastore.unionmetadatastore( | |
523 | *repo.sharedhistorystores) |
|
523 | *repo.sharedhistorystores) | |
524 |
|
524 | |||
525 | missingids = set() |
|
525 | missingids = set() | |
526 | if fetchdata: |
|
526 | if fetchdata: | |
527 | missingids.update(datastore.getmissing(idstocheck)) |
|
527 | missingids.update(datastore.getmissing(idstocheck)) | |
528 | if fetchhistory: |
|
528 | if fetchhistory: | |
529 | missingids.update(historystore.getmissing(idstocheck)) |
|
529 | missingids.update(historystore.getmissing(idstocheck)) | |
530 |
|
530 | |||
531 | # partition missing nodes into nullid and not-nullid so we can |
|
531 | # partition missing nodes into nullid and not-nullid so we can | |
532 | # warn about this filtering potentially shadowing bugs. |
|
532 | # warn about this filtering potentially shadowing bugs. | |
533 | nullids = len([None for unused, id in missingids if id == nullid]) |
|
533 | nullids = len([None for unused, id in missingids if id == nullid]) | |
534 | if nullids: |
|
534 | if nullids: | |
535 | missingids = [(f, id) for f, id in missingids if id != nullid] |
|
535 | missingids = [(f, id) for f, id in missingids if id != nullid] | |
536 | repo.ui.develwarn( |
|
536 | repo.ui.develwarn( | |
537 | ('remotefilelog not fetching %d null revs' |
|
537 | ('remotefilelog not fetching %d null revs' | |
538 | ' - this is likely hiding bugs' % nullids), |
|
538 | ' - this is likely hiding bugs' % nullids), | |
539 | config='remotefilelog-ext') |
|
539 | config='remotefilelog-ext') | |
540 | if missingids: |
|
540 | if missingids: | |
541 | global fetches, fetched, fetchcost |
|
541 | global fetches, fetched, fetchcost | |
542 | fetches += 1 |
|
542 | fetches += 1 | |
543 |
|
543 | |||
544 | # We want to be able to detect excess individual file downloads, so |
|
544 | # We want to be able to detect excess individual file downloads, so | |
545 | # let's log that information for debugging. |
|
545 | # let's log that information for debugging. | |
546 | if fetches >= 15 and fetches < 18: |
|
546 | if fetches >= 15 and fetches < 18: | |
547 | if fetches == 15: |
|
547 | if fetches == 15: | |
548 | fetchwarning = self.ui.config('remotefilelog', |
|
548 | fetchwarning = self.ui.config('remotefilelog', | |
549 | 'fetchwarning') |
|
549 | 'fetchwarning') | |
550 | if fetchwarning: |
|
550 | if fetchwarning: | |
551 | self.ui.warn(fetchwarning + '\n') |
|
551 | self.ui.warn(fetchwarning + '\n') | |
552 | self.logstacktrace() |
|
552 | self.logstacktrace() | |
553 | missingids = [(file, hex(id)) for file, id in missingids] |
|
553 | missingids = [(file, hex(id)) for file, id in missingids] | |
554 | fetched += len(missingids) |
|
554 | fetched += len(missingids) | |
555 | start = time.time() |
|
555 | start = time.time() | |
556 | missingids = self.request(missingids) |
|
556 | missingids = self.request(missingids) | |
557 | if missingids: |
|
557 | if missingids: | |
558 | raise error.Abort(_("unable to download %d files") % |
|
558 | raise error.Abort(_("unable to download %d files") % | |
559 | len(missingids)) |
|
559 | len(missingids)) | |
560 | fetchcost += time.time() - start |
|
560 | fetchcost += time.time() - start | |
561 | self._lfsprefetch(fileids) |
|
561 | self._lfsprefetch(fileids) | |
562 |
|
562 | |||
563 | def _lfsprefetch(self, fileids): |
|
563 | def _lfsprefetch(self, fileids): | |
564 | if not _lfsmod or not util.safehasattr( |
|
564 | if not _lfsmod or not util.safehasattr( | |
565 | self.repo.svfs, 'lfslocalblobstore'): |
|
565 | self.repo.svfs, 'lfslocalblobstore'): | |
566 | return |
|
566 | return | |
567 | if not _lfsmod.wrapper.candownload(self.repo): |
|
567 | if not _lfsmod.wrapper.candownload(self.repo): | |
568 | return |
|
568 | return | |
569 | pointers = [] |
|
569 | pointers = [] | |
570 | store = self.repo.svfs.lfslocalblobstore |
|
570 | store = self.repo.svfs.lfslocalblobstore | |
571 | for file, id in fileids: |
|
571 | for file, id in fileids: | |
572 | node = bin(id) |
|
572 | node = bin(id) | |
573 | rlog = self.repo.file(file) |
|
573 | rlog = self.repo.file(file) | |
574 | if rlog.flags(node) & revlog.REVIDX_EXTSTORED: |
|
574 | if rlog.flags(node) & revlog.REVIDX_EXTSTORED: | |
575 | text = rlog.revision(node, raw=True) |
|
575 | text = rlog.revision(node, raw=True) | |
576 | p = _lfsmod.pointer.deserialize(text) |
|
576 | p = _lfsmod.pointer.deserialize(text) | |
577 | oid = p.oid() |
|
577 | oid = p.oid() | |
578 | if not store.has(oid): |
|
578 | if not store.has(oid): | |
579 | pointers.append(p) |
|
579 | pointers.append(p) | |
580 | if len(pointers) > 0: |
|
580 | if len(pointers) > 0: | |
581 | self.repo.svfs.lfsremoteblobstore.readbatch(pointers, store) |
|
581 | self.repo.svfs.lfsremoteblobstore.readbatch(pointers, store) | |
582 | assert all(store.has(p.oid()) for p in pointers) |
|
582 | assert all(store.has(p.oid()) for p in pointers) | |
583 |
|
583 | |||
584 | def logstacktrace(self): |
|
584 | def logstacktrace(self): | |
585 | import traceback |
|
585 | import traceback | |
586 | self.ui.log('remotefilelog', 'excess remotefilelog fetching:\n%s\n', |
|
586 | self.ui.log('remotefilelog', 'excess remotefilelog fetching:\n%s\n', | |
587 | ''.join(traceback.format_stack())) |
|
587 | ''.join(traceback.format_stack())) |
General Comments 0
You need to be logged in to leave comments.
Login now