1 2008-08-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
3 * tempted to refactor rmirror into resolve_symlink, localize, etc.
4 Curious if rsync_options=links equal 0 vs. 1 will make the expected
7 * rsync options: it's a bit of a pain that we usually need several rsync
8 options, like compress, links, times, checksum and that there is no
9 reasonable default except the original rsync default. I think wee can
10 afely assume that the rsync options are shared between all recentfile
11 instances within one recent tree.
13 2008-08-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
15 * deletes: if a delete follows an add quickly enough it may happen that
16 a downstream mirror did not see the add at all! It seems this needs to
17 be mentioned somewhere.
19 2008-08-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
21 * I suspect the treat of metadata is incorrect during read or something.
22 The bug that I am watching is that between 06:08 and 06:09 the 6h file
23 contained more than 6 hours worth of data. At 06:08 we merged into the
24 1d file. We need to take snapshots of the 6h file over the course of an
25 hour or maybe only between XX:08 and XX:09? Nope, the latter is not
28 Much worse: watching the 1h file: right at the moment (at 06:35) it
29 covers 1218867584-1219120397 which is 70 hours.
31 Something terribly broken. BTW, 1218867584 corresponds to Sat Aug 16
32 08:19:44 2008, that is when I checked out last time, so it seems to be
33 aggregating and never truncating?
35 No, correct is: it is never truncating; but wrong is: it is aggregating.
36 It does receive a lot of events from time to time from a larger file.
37 Somehow a large file gets merged into the small one and because the
38 "meta/merged" attribute is missing, nobody is paying attention. I
39 believe that I can fix this by making sure that metadata are honoured
40 during read. DONE and test adjusted.
42 2008-08-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
46 remotebase => remoteroot to fit well with localroot DONE
47 local_path() => localroot seems to me should already work DONE
48 recentfile_basename => rfilename no need to stress it has no slash DONE
50 filenameroot??? Doesn't seem too bad to me today. Maybe something like
51 kern? It would anyway need a deprecation cycle because it is an
52 important constructor.
54 * I like the portability that Data::Serializer brings us but the price
55 is that some day we might find out that it is slowing us a bit. We'll
58 2008-08-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
60 * should we not enter the interval of the principal (or the interval of
61 the merging file?) in every aggregated/merged file?
63 * we should aim at a first release and give up on thinking about
64 sanitizing stuff and zloop. Let's just admit that a full traditional
65 rsync is the only available sanitizer ATM. Otherwise it's complicated
66 stuff: sanitizing on the origin server, sanitizing on the slaves,
67 sanitizing forgotten files, broken timestamps, etc. Let's delay it and
68 get the basics out before this becomes a major cause for mess.
70 2008-08-13 Andreas Koenig <k@andreas-koenigs-computer.local>
72 * On OSes not supporting symlinks we expect that RECENT.recent contains
73 the contents of the principal recentfile. Actually this is identical on
74 systems supporting symlinks. Simple, what follows from that is that we
75 need to keep the serializer in the metadata because we cannot read it
76 from the filename, doesn't it? Of course not. It's a chicken and egg
77 problem. This leaves us with the problem to actually parse the
78 serialized data to find out in which format it is. So who can do the 4
79 or 5 magics we wanted to support? File::LibMagic?
81 2008-08-09 Andreas Koenig <k@andreas-koenigs-computer.local>
83 * remotebase and recentfile_basename are ugly names. Now that we need a
84 word for the shortest/principal/driving recentfile too we should do
87 localroot is good. rfile is good. local_path() is bad, local_path($path)
88 is medium, filenameroot() is bad, remotebase is bad, recentfile is
91 Up to now remotebase was the string that described the remote root
92 directory in rsync notation, like pause.perl.org::authors. And
93 recentfile_basename was "RECENT-1h.yaml".
95 2008-08-08 Andreas Koenig <k@andreas-koenigs-computer.local>
97 * The test that was added in today's checkin is a good start for a test
98 of rmirror. We should have more methods in Recent.pm: verify,
99 addmissingfiles. We should verify the current tree, then rmirror it and
100 then verifytree the copy. We could then add some arbitrary file and let
101 it be discovered by addmissingfiles, then rmirror again and then
102 verifytree the copy again.
104 Then we could start stealing from csync2 sqlite database [no port to
105 OSX!] and fill a local DB. And methods to compare the database with the
106 recentfiles. Our strength is that in principle we could maintain state
107 with a single float. We have synced up to 1234567890.123456. If the Z
108 file does not add new files all we have to do is mirror the new ones and
111 This makes it clear that we should extend current protocol and declare
112 that we cheat when we add files too late, just to help the other end
113 keeping track. Ah yes, that's what was meant when zloop was mentioned
116 Maybe need to revisit File::Mirror to help me with this task.
118 2008-08-07 Andreas Koenig <k@andreas-koenigs-computer.local>
120 * There must be an allow-me-to-truncate flag in every recentfile.
121 Without it one could construct a sequence of updates winning the locking
122 battle against the aggregator. Only if an aggregator has managed to
123 merge data over to the next level, truncating can be allowed. DONE with
126 2008-08-06 Andreas Koenig <k@andreas-koenigs-computer.local>
128 * We should probably guarantee that no duplicates enter the aggregator
131 2008-08-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
133 * To get merge operation faster would need a good benchmark test. What
134 02 spits out isn't reliable enough and is dominated by many other
137 commit 10176bf6b79865d4fe9f46e3857a3b8669fa7961
138 Author: Andreas J. Koenig <k@k75.(none)>
139 Date: Sat Aug 2 07:58:04 2008 +0200
143 commit 3243120a0c120aaddcd9b1f4db6689ff12ed2523
144 Author: Andreas J. Koenig <k@k75.(none)>
145 Date: Sat Aug 2 11:40:29 2008 +0200
147 there was a lot of trying but the effect is hardly measurable with
150 * overhead of connecting seems high. When setting
151 max_files_per_connection to 1 we see that.
153 2008-08-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
155 * 1217622571.0889 - 1217597432.86734 = 25138.2215600014
157 25138.2215600014/3600 = 6.98283932222261
159 It jumps into the eye that this is ~ 7 hours, not ~6, so there seems to
160 be a bug in the aggregator. FIXED
162 2008-07-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
164 * e.g. id/Y/YE/YEWENBIN/Emacs-PDE-0.2.16.tar.gz: Do we have it, should
165 we have it, can we mirror it, mirror it!
167 I fear this needs a new class which might be called
168 File::Rsync::Mirror::Recent. It would collect all recentfiles of a kind
169 and treat them as an entity. I realize that a single recentfile may be
170 sufficient for certain tasks and that it is handy for the low level
171 programmer but it is not nice to use. If there is a delete in the 1h
172 file then the 6h file still contains it. Seekers of the best information
173 need to combine at least some of the recentfiles most of the time.
175 There is the place for the Z loop!
177 But the combination is something to collect in a database, isn't it. Did
178 csync2 just harrumph?
180 2008-07-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
182 * it just occurred to me that hosts in the same mirroring pool could
183 help out each other even without rewriting the recentfile. Just fetch
184 the stuff to mirror from several places, bingo. But that's something
185 that should rather live in a separate package or in rsync directly.
187 * cronjobs are unsuited because with ntp they would all come at the full
188 minute and disturb each other. Besides that I'd hate to have a backbone
189 with more than a few seconds latency.
191 2008-07-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
193 * a second rsync server with access control for PAUSE. Port? 873 is the
194 standard port, let's take 8873.
196 * if there were a filesystem based on this, it would have a slow access
197 to inexistent files. It would probably provide wrong readdir (only based
198 on current content) or also a slow one (based on a recentfile written
199 after the call). But it would provide fast access to existing files. Or
200 one would deliberately allow slightly blurred answers based on some
201 sqlite reflection of the recentfiles.
203 * todo: write a variant of mirror() that combines two or more
204 recentfiles and treats them like one
206 * todo: signal handler to remove the tempfile
208 2008-07-24 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
210 * now that we have the symlink I forgot how it should be used in
213 * the z loop: add missing files to Z file. Just append them (instead of
214 prepending). So one guy prepends something from the Y file from time to
215 time and another guy appends something rather frequently. Collecting
216 pond. When Y merges into Z, things get epoch and the collecting pond
217 gets smaller. What exactly are "missing files"?
219 take note of current epoch of the alpha file, let's call it the
222 find all files on disk
224 remove all files registered in the recentworld up to recent-ts
226 remove all files that have been deleted after recent-ts according to
229 2008-07-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
231 * rersyncrecent might be a cronjob with a (locked) state file which
232 contains things like after and maybe last z sync or such?
234 rrr-mirror might be an alternative name but how would we justify the
235 three Rs when there is no Re-Rsync-Recent?
237 With the --loop parameter it is an endless loop, without it is no loop.
238 At least this is simple.
240 * todo: new accssor z-interval specifies how often the Z file is updated
241 against the filesystem. We probably want no epoch stamp on these
242 entries. And we want to be able to filter the entries (e.g. no
243 by-modules and by-category tree)
245 2008-07-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
247 * Fill the Z file. gc or fsck or both. Somehow we must get the old files
248 into Z. We do not need the other files filled up with filesystem
251 * need interface to query for a file in order to NOT call update on
252 PAUSE a second time within a short time.
254 2008-07-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
256 * recommended update interval? Makes no sense, is different for
263 change-log-default-name: "Todo"